Modeling Spongy Moth Forest Mortality in Rhode Island Temperate Deciduous Forest

Modeling Spongy Moth Forest Mortality in Rhode Island Temperate Deciduous Forest | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Modeling Spongy Moth Forest Mortality in Rhode Island Temperate Deciduous Forest Liubov Dumarevskaya, Jason R. Parent This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-4378454/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Invasive pests cause major ecological and economic damages to forests around the world. Satellite imagery is an important tool for monitoring defoliation at broad scales, but environmental conditions can affect whether defoliation leads to mortality. In this study, we modeled forest mortality resulting from a 2015-2017 Spongy moth outbreak in the temperate deciduous forests of Rhode Island (northeastern U.S.). We used Landsat-based defoliation mapping and geospatial environmental data with Random Forest to model mortality severity of canopy trees at a 100 m spatial resolution. Defoliation was mapped based on declines in Normalized Difference Vegetation Index (NDVI) in outbreak years compared to baseline NDVI from pre-outbreak years. Other predictors included geospatial data representing soil characteristics, drought condition, and forest characteristics as well as proximity to coast, development, and water. The Random Forest tool in Python Sklearn was used to model forest mortality with 2 classes (low/high) and 3 classes (low/med/high). The best models had overall accuracies of 82% and 65% for the 2-class and 3-class models, respectively. The most important predictors of forest mortality were defoliation, distance to coast, and canopy cover. Soils and topography had minimal importance in the models possibly due to limitations of the data and limited variability within our study area. Repeated defoliations were relatively rare during the outbreak. Model performance improved only slightly with the inclusion of more than 3 variables. The models classified 35% of forests as having canopy mortality > 5 trees/ha and 21% of Rhode Island forests having mortality >11 trees/ha. The study shows the benefit of Random Forest models that use both defoliation maps and geospatial environmental data for classifying forest mortality. Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 1. Introduction Invasive pests have far-reaching ecological and economic impacts, affecting global forest communities and carbon storage (Seidl et al., 2014). Globally, Quirion et al. (2021) found that forests affected by pest invasion sequester 69% less carbon, on average, compared to unaffected forests. In the U.S. alone, forest pest activity results in annual biomass losses of 5.5 TgC (Fei et al., 2019), which is roughly equivalent to 11,800 hectares of temperate forest land (based on Murray et al., 2011). One important forest pest of the U.S. is Spongy moth ( Lymantria dispar dispar ) which can cause severe tree defoliation that leads to mortality. From 1994–2010, Spongy moth caused an estimated 898 TgC of biomass loss in the U.S. (Fei et al., 2019), which is equivalent to roughly 1.9 M hectares of temperate forest (Murray et al., 2011). In a single U.S. city (i.e. Baltimore, MD), the damages caused by Spongy moth was estimated to range from $ 5.5–63.7 M per year of outbreak, depending on environmental and management scenarios (Bigsby et al., 2014). Satellite and aerial imagery allow for a regular monitoring of forest health and tree mortality (Verbesselt et al., 2009). Sentinel-2 and Landsat satellites provide moderate-resolution imagery over large areas. Summertime aerial imagery is acquired regularly by various agencies including the USDA’s National Agriculture Inventory Program (NAIP). Thus, remote sensing provides an important tool for monitoring forest health which can allow foresters to mitigate the impact of insect invasions through forest management (Hudgins et al., 2017). Mitigation can include direct prevention strategies such as biological controls and insecticides; indirect controls such as forest thinning, prescribed fire, and quarantines (Ivantsova et al., 2019); or replanting with pest-resistant varieties of the affected species (Kinahan et al., 2020). Tree mortality maps can guide tree removal programs to protect public safety and infrastructure, maintain aesthetics (Poulos et al., 2010; Guggenmoose et al., 2003), improve forest regeneration (Parotta et al., 1997), or convert forests with high mortality to early successional wildlife habitat. Defoliation by Spongy moth and other leaf-eating insects, has a strong impact on tree mortality (Gottschalk et al., 2007; Baker et al., 1941). Davidson et al. (1999) reported that Spongy moth invasion caused a die-off of up to 47% of deciduous trees in outbreak areas in the Eastern U.S. When more than 75% of a tree crown is defoliated, a tree may produce a second set of leaves in the same growing season which depletes energy reserves and decreases chances of surviving the invasion (Davidson et al., 1999), especially if the tree is defoliated multiple times (Campbell, 1979). Other pests, such as borers, are attracted to stressed trees which further increases chances of tree mortality (Davidson 1999). Researchers have mapped defoliation, based on satellite and aerial imagery, and used these maps to model tree mortality (Meddens et al., 2014; Goodwin et al., 2008; Long et al., 2016; Bergmüller et al., 2022; Meng et al., 2022; Zhan et al., 2020). In the Rocky Mountains, Meddens et al. (2014) mapped dead pine trees with Landsat-based vegetation indexes and linear regression models (r 2 = 0.77). In boreal forests, Goodwin et al. (2008) classified dead pine trees, with 75% accuracy, using Normalized Difference Moisture Index derived from Landsat imagery. In spruce-dominated forests, Long et al. (2016) classified dead trees, with 90% accuracy, using a combination of high-resolution aerial imagery and moderate-resolution satellite imagery. In the forests of western Canada, Bergmuller et al. (2022) used multispectral drone imagery to map tree mortality with approximately 70% and 80% accuracy for deciduous and coniferous trees, respectively. In China, Meng et al. (2022) used multispectral satellite imagery to map Southern Pine Beetle infestations with an accuracy > 80%. In pine forests in northern China, Zhan et al. (2020) used GF-2 and Sentinel-2 imagery to map tree mortality with 78% accuracy. Pasquarella et al. (2018) mapped defoliation from a Spongy moth outbreak with Landsat imagery using a tasseled-cap greenness index, but they did not model actual mortality. Defoliation is a driving factor in tree mortality which can be compounded by various environmental stressors. Soil and geology characteristics may directly impact water availability, root depth, and nutrient availability, which can stress trees and decrease survival during defoliation events. Topographic position (i.e. hilltop, mid-slope or valley) can affect light and precipitation distribution, ground surface temperature, soil conditions, and depth of the groundwater which collectively determine water availability (Zhang et al., 2022; Øystein et al, 2015; Dunn et al., 1987). When analyzing factors related to the cause of death of individual trees in tropical moist forests of northern Amazonia, Toledo (2011) found that soil characteristics and topographic relief together accounted for 20% of the mortality variation. They also found that tree mortality was higher on steep slopes and on sandy soils in valleys while mortality was lower on plateaus with clay soils. Drought conditions reduce water availability and adds stress to defoliated tree (Bottero et al., 2017, Ramsfield et al., 2016; Choat et al., 2018). Guarin et al. (2005) found correlations between multiple years of drought and tree mortality. Tree in close proximity to roads and urban areas may experience greater stress and mortality due to exposure to de-icing salt (Kayama et al., 2003, Horsley et al., 2002) and air pollution (Trumbore et al., 2015, Percy et al., 2004). Forests near urban areas may have higher exposure to invasive plants or pests (Stravinskienė et al., 2018). In New England (U.S.), Davidson et al. (1999) reported that overall tree health significantly affected survival from Spongy moth outbreaks with mortality rates of 36% and 7% for unhealthy and healthy trees, respectively. Forest characteristics can influence tree mortality during pest outbreaks. Some invasive pests have species-specific feeding preferences and, thus, preferred species tend to be more heavily affected (Davidson et al., 1999; Barbosa et al., 1978). Canopy cover has been found to be an important factor in mortality prediction, although the effect was not consistent across different forest types and environments (Campbell et al., 2020, Gunst et al., 2016, Dorman et al., 2015, Das et al., 2011). In woodland areas of Utah, Campbell et al. (2020) found that high mortality is more likely to occur in areas with low-to-moderate canopy cover (5–25%). Gunst et al. (2016) found that higher canopy cover was positively associated with tree mortality in xeric pine forests but negatively associated with tree mortality in red fir forests, during wet years. Das et al. (2011) found that canopy cover did not have a major effect on mortality in forests in the western U.S. where mortality was driven by insects and diseases. Dorman et al. (2015) found no association between canopy cover and mortality in pine forests. A few studies have included both tree mortality and environmental geospatial data as predictors in tree mortality models. In open canopy semi-arid woodlands, Campbell et al. (2020) found that the inclusion of environmental factors significantly improved predictions of tree mortality when combined with defoliation mapped from high-resolution aerial imagery and Landsat imagery. Navarro-Cerrillo et al. (2019) mapped dead holm oak trees ( Quercus ilex ) with 87% accuracy using WorldView-2 satellite imagery and light detection and ranging (LiDAR) data. They identified a link between high mortality and specific soil properties in Spanish woodlands impacted by root rot decline. Most tree mortality models have been based solely on defoliation metrics and they tended to be conducted in coniferous forests which were dominated by few tree species (Meddens et al., 2013; Meddens et al., 2014; Bergmüller et al., 2022; Goodwin et al., 2008; Shearman et al., 2019). Studies that modeled tree mortality for mixed deciduous forests, with a wide variety of species, tended use very high-resolution drone imagery (Bergmüller et al., 2022, Long et al., 2016), which is not practical to collect for large study areas. Some studies have included both environmental factors and satellite-based defoliation maps in tree mortality models (Campbell et al. 2020, Navarro-Cerrillo et al., 2019); however, we are not aware of any studies that have applied these methods in temperate deciduous forests. The objective of this paper is to model tree mortality resulting from a 2015–2017 Spongy moth outbreak in the temperate deciduous forest in Rhode Island, located in the northeastern U.S. We use a random forest approach to model mortality based on defoliation and environmental factors. We map defoliation from Landsat imagery and include geospatial data representing topography, climate, soil, and vegetation characteristics in modeling mortality. This research explores the importance of defoliation and environmental factors, represented by geospatial data, in predicting tree mortality from forest pest outbreaks. 2. Methods 2.1. Study Area Our study area was the state of Rhode Island which has a total area of 3,144 km 2 . Rhode Island is largely covered by temperate deciduous forests dominated by oak ( Quercus sp.) and hickory ( Carya sp.) species (Enser et al., 2011). Other common species include white pine ( Pinus strobus ), sugar maple ( Acer saccharum ), red maple ( Acer rubrum ), American beech ( Fagus grandifolia ), hemlock ( Tsuga canadensis ), birch ( Betula sp .), elm ( Ulmus americana ), white ash ( Fraxinus americana) , and linden tree (Tilia americana) (Enser et al., 2011). Coniferous forests of Rhode Island are dominated by eastern white pine ( Pinus strobus ) and occur mostly in the southern part of state (Enser et al., 2011). The topography of Rhode Island mostly consists of gently rolling uplands with elevations up to 60 m and hillier upland terrain in the northwest with elevations of 60 to 200 m. The state includes humid continental, humid subtropical, and oceanic climate types (NCDC, 1971). Mean annual temperatures in Rhode Island range from 8-12 o C and annual precipitation ranges from 100–150 cm with lower precipitation amounts in coastal areas. Summers are characterized by periods (i.e. > 2 weeks) of little to no precipitation. Rhode Island is the second most densely populated state in the U.S. with an average of 393 persons / km 2 (USCB, 2020). The state contains 1784 km 2 of forest land which generates an estimated $ 408 M in forest products and $ 720 M annually in recreational value (RIDEM, 2019). In addition, the forests provide erosion and flood control and other ecosystem services (RIDEM, 2019). The study area experienced a severe spongy moth outbreak in 2015–2017. This outbreak defoliated approximately 4386 km 2 of forest in New England which made it the largest outbreak in 30 years (Pasquarella et al., 2018). This outbreak was caused by unusually dry springs in 2014, 2015, and 2016. Those dry conditions suppressed a growth of fungus Entomophaga maimaiga which typically controls spongy moth populations in the region (Pasquarella et al., 2018). 2.2 Data and Software To map defoliation, we used atmospherically corrected (level 2) Landsat 8 satellite imagery acquired in mid-summer (i.e. late June to early August) from 2009–2017 (USGS, 2022). Landsat imagery has a 30 m spatial resolution for the visible and near-infrared wavelengths. Images corresponded as closely as possible to early July after spongy moth activity reached its peak and before significant refoliation could occur. We acquired cloud-free imagery for 2009, 2011, 2013, and 2016. No cloud-free images were available for the remaining years so we to used multiple images to create cloud-free composites. The composite images were created by using a cloud mask layer to identify cloud pixels in the primary images and replacing those pixels with cloud-free image pixels from the secondary images. Secondary images had acquisition dates within 4 weeks of the primary images. We used high-resolution summertime aerial imagery for July 2016 (RIGIS, 2023) to evaluate defoliation mapping for the corresponding year. The imagery had a spatial resolution of 7.6 cm and includes visible and near-infrared wavelengths. To evaluate defoliation mapping for 2015 and 2017, we used atmospherically corrected Sentinel-2 imagery (USGS, 2022). Sentinel-2 imagery has a spatial resolution of 10 m for the visible and near-infrared wavelengths. To create training and validation for the tree mortality models, we used aerial imagery acquired during in leaf-on summer period in July-August of 2019 with visible (RGB) bands and a 7.6 cm spatial resolution (RIGIS, 2023). Various datasets were used to create environmental metrics. Most of these datasets were downloaded from the Rhode Island Geographic Information System and included Soils , Glacial Deposits , Ecological Communities of Rhode Island , 2011 Rhode Island Statewide LiDAR, Freshwater Lakes, Ponds and Reservoirs , and Coastal Waters (RIGIS, 2023). We also used the U.S. Drought Monitor data (NDMC, 2022) and a 1 m resolution land cover from the National Oceanographic and Atmospheric Administration Coastal Change Analysis Program (CCAP) (NOAA, 2016). We used weekly Drought Monitor data for the summertime (June 1 – Sept. 1) during the spongy moth outbreak (2015–2017). The LiDAR data were acquired during leaf-off conditions in March and had a point density of 2 + pts/m 2 . The ground-classified points in the LiDAR data were used to create a 1-m resolution digital elevation model (DEM). For all the GIS-based data processing and analysis we used ArcGIS Pro with Python 3.11.3. The Python modules SciKit-Learn, Seaborn, and SHAP were used to run Random Forest analysis, calculate related metrics, and create figures for tree mortality modelling. 2.3. Predictors of Tree Mortality We created a 100 m x 100 m (1 ha) polygon grid covering the study area to serve as the analysis units for modeling. Tiles were included in the analysis if they had at least 75% forest cover (i.e. deciduous, coniferous, mixed, wetland), based on the CCAP land cover data. Most of the environmental metrics represent the fraction of a given tile that corresponds to the particular feature-of-interest. Altogether, we developed 21 environmental metrics which characterized soil properties, topography, climate, forest characteristics, and proximity to resources and stressors (Table 1 ). 2.3.1. Defoliation We used Normalized Difference Vegetation Index (NDVI) to map defoliation with 30 m spatial resolution based on Landsat imagery (Rullan-Silva et al., 2013). To establish a baseline NDVI for the pre-outbreak forest, we calculated NDVI for the 2009, 2011, and 2013 summer images. We estimated annual pre-outbreak variation in NDVI based on a sample of 200 points. The sample included 100 points in locations with obvious defoliation and 100 points in locations with no apparent defoliation, based on the 2016 summertime aerial imagery. Points were semi-randomly distributed to encompass the extent of forest land in Rhode Island. The NDVI values for the pixels corresponding to the sample point locations were extracted from the pre-outbreak imagery. We calculated the average pre-outbreak NDVI values for each pixel as well as the standard deviations for each year. For each outbreak year (2015–2017), the NDVI rasters were subtracted from the pre-outbreak average NDVI raster. We considered NDVI differences of > 0.09 to correspond to significant defoliation. This threshold was double the largest pre-outbreak standard deviation which helped ensure that the differences exceeded the typical yearly variations. We calculated a defoliation index for the 100 m tiles which weighted multiple years of defoliation more heavily than a single year of defoliation. The defoliation index was calculated as: $$Defoliation Index=fD1 + 2*fD2 + 3*fD3$$ 1 where fD1 , fD2 , and fD3 are the fraction of area within each tile that experienced defoliation for 1, 2, or 3 years respectively (Table 1 ). 2.3.2. Environmental metrics Canopy cover can have positive or negative impacts on tree mortality depending on the ecosystem. We used the LiDAR point cloud to calculate canopy cover based on the first-return cover index (FRCI) (Ma et al., 2017). This metric was calculated, based on the LiDAR points within each tile, as the ratio of first returns that intercept the canopy (First Canopy ) to the total number of first returns(First Total ): \(\) $$FRCI={\Sigma }\text{F}\text{i}\text{r}\text{s}\text{t}\text{C}\text{a}\text{n}\text{o}\text{p}\text{y} / {\Sigma }\text{F}\text{i}\text{r}\text{s}\text{t}\text{T}\text{o}\text{t}\text{a}\text{l}$$ 2 First returns indicate the location where the LiDAR pulse first intercepted an object. We considered first returns with heights > 3 m to correspond to the canopy. FRCI values range from 0 (no cover) to 1 (complete cover) (Table 1 ). The amount of coniferous tree cover can be important factor because Spongy moth prefers to feed on deciduous trees. Also, ruderal and plantation forests have different tolerances for stressors than later successional forests or forests that are not heavily managed. Thus, we calculated the fractions of each tile covered by evergreen, ruderal, and plantation forests, based the Ecological Communities dataset (Table 1 ). Drought conditions can have a major impact on tree survival when it coincides with other stressors such as defoliation. We calculated a drought index during the peak growing season (i.e. June through August) for the 2015–2017 outbreak period (Table 2 ). The drought data consisted of polygons with attributes ranging from 0–6 to indicate conditions ranging from no drought to severe drought (NDMC 2022). The polygons were intersected with the 100 m grid to join the weekly drought indices to our study tiles. The drought index was the average of the weekly drought ratings over all the summer seasons of the outbreak period for each tile. Coastal environments have differing climatic conditions compared to inland environments. To represent coast proximity , we calculated distances from the coastline as defined by the Rhode Island Coastal Waters dataset. The distances within each tile were averaged to create the metric (Table 2 ). Urban proximity can stress trees through exposure to air pollution, road salt, heat island effects, and invasive pests (Stravinskiene et al., 2018, Trumbore et al., 2015, Kayama et al., 2003, Percy et al., 2004, Horsley et al., 2002). We calculated urban distance as distance from developed land cover in the CCAP land cover dataset. The distances within each tile were averaged to create the metric (Table 2 ). Proximity to waterbodies can influence soil moisture levels, nutrient availability, and the prevalence of certain pests and diseases. We used the Freshwater Lakes, Ponds and Reservoirs dataset to calculate distance from waterbodies. The distances within each tile were averaged to create the Lake proximity metric (Table 2 ). Soil and glacial deposits may be limiting to root growth or water availability (Table 3 ). Soil attributes included tile fractional cover of 1) hydric soils with permanent or temporary oversaturation, 2) overdrained soils, 3) eroded soils, 4) densic horizon close to the surface, 5) stony soils, and 6) restrictive soils. Metrics relating to glacial deposit attributes included tile fractional cover of 1) till, 2) outwash, and 3) bedrock. Topography can influence climate which may affect tree survival during stressful events. Hilltops and steep south-facing slopes may be drier than valleys or north-facing slopes. To characterize relevant slope conditions, we calculated slope and aspect based the 1 m DEM. Slope and aspect corresponded to the steepest gradient within a 3x3 window centered on each pixel (Esri, 2022). We extracted slopes > 20° as well as slopes that were both steep (> 20°) and south-facing with aspects between 225° and 315°. Metrics were calculated as the fraction of the tiles covered by steep slopes and steep south-facing slopes (Table 3 ). To identify hilltops and valley bottoms, we used a Topographic Position Index (TPI) (Reu et al., 2013) based on the DEM. We calculated TPI by subtracting the average elevation within 20 m of a given pixel location from the actual elevation at that location. Positive TPI values indicate hilltops whereas negative values indicate valley bottoms. We extracted TPI > 0.2 and TPI < 0.2 to represent hilltops and valley bottoms, respectively. The fractional cover of tiles covered by hilltops and valley bottoms were used as the metrics (Table 3 ). Table 1 List of forest-related predictors. Name Description Justification Data Defoliation index % defoliated forest cover weighted by number of years of defoliation. Defoliation depletes energy reserves. Landsat Evergreen cover % cover by coniferous trees Coniferous trees are not a preferred food source but do not tolerate defoliation. Ecological Communities Canopy cover % of 1st returns intercepted by canopy Canopy cover can affect competition and microclimate. 2011 Lidar Ruderal forests % cover by early successional forest species Early successional forest has species adapted to disturbances. Heavily managed forest may have different tolerances for stressors. Ecological Communities Plantations % cover plantation forest Table 2 List of proximity- and drought-related predictors. Name Description Justification Data Drought index Average summer drought rating during outbreak Drought is a major stressor. U.S. Drought Monitor Coast proximity Average distance to the coast Proximity to coast affects climate conditions. RI Coastal Waters Urban proximity Average distance to development Developed areas may be exposed to more pollutants and invasive species. NOAA CCAP land cover Lake proximity Average distance to the lakes Locations near water bodies may be associated with more abundant groundwater. Lakes Ponds and Reservoirs Table 3 List of soil- and topography-related predictors. Name Description Justification Data Hydric soils % cover by hydric soils Hydric soils may have greater water availability whereas overdrained soils may have low available moisture. Restrictive soils, shallow bedrock, tills, and stony soils can limit root depth or lateral distribution, thus, reducing access to water and nutrients. Eroded and outwash soils may have low nutrient availability. NRCS Soils Restrictive soils % cover by restrictive soils Stony soils % cover by stony soils Shallow bedrock % cover by soils with shallow bedrock Overdrained soils % cover by overdrained soils Eroded soils % cover by eroded soils Outwash % cover by glacial outwash Glacial deposits Till % cover by glacial till Valley bottoms % cover by valleys (TPI 0.2) Steep slopes % cover by slopes > 20 \(^\circ\) Steeper slopes can affect soil moisture, nutrient availability, and stability. South-facing tend to be warmer and drier. Steep south-facing slopes % steep slopes facing 225°-315° 2.4. Training/validation data To train and validate our tree mortality models, we used two different “ground-truth” datasets that were based on summertime 2019 aerial imagery. The first set included 1426 tiles that were randomly selected from the 100 m polygon grid. These tiles were visually evaluated using the aerial imagery and rated using a six-category system to characterize mortality of overstory trees in each tile (Table 4 ). The second training/validation dataset included 787 tiles corresponding to Rhode Island Department of Environmental Management (RIDEM) property that had particularly high rates of tree mortality. For these tiles, the locations of dead overstory trees were digitized as point features and subsequently aggregated by tile to create ratings that were consistent with the first training dataset (Table 4 ). Altogether, the training/validation dataset included 2320 sample tiles. Table 4 Classes of tree mortality in training/validation dataset. Class Dead trees / ha Number of Tiles (Dataset 1) Number of Tiles (Dataset 2) Total Number of Tiles 0 0 258 20 278 1 1–2 299 50 349 2 3–5 239 101 340 3 6–10 222 176 398 4 11–20 152 272 424 5 > 20 8 523 531 2.5. Modeling Tree Mortality using Random Forest We used Random Forest (Scikit-Learn, 2022) to model tree mortality based on the 21 environmental and defoliation metrics (Tables 1 – 3 ). We created two different models – one predicted two classes of mortality (low/high) and the other predicted three classes of mortality (low/med/high). The two-class model predicted low and high mortality classes with 11 dead overstory trees per hectare. We selected these thresholds to ensure there were relatively similar numbers of tiles representing each class. We used 5-fold cross-validation with a split of 80% and 20% for training and validation data, respectively, for each fold. We tuned our model by adjusting several hyperparameters through trial and error to find optimal parameters. Hyperparameters are the settings that are not learned by the model during training but set beforehand, and they can significantly impact the model's performance. We adjusted the following hyperparameters: number of trees, minimum number of samples required to be at a leaf node, and the criterion for splitting a node. We found the models performed best with 100 trees, a minimum of 3 features per node, and using the Gini criterion for splitting the nodes. 2.5. Accuracy Assessment We assessed the accuracy of our defoliation mapping for 2015, 2016 and 2017 by qualitatively assessing Landsat, aerial, and Sentinel imagery, respectively. These image datasets represented the best available data for each year. For both defoliation mapping and the mortality models, we quantitatively assessed accuracy using F1 score, precision, and recall. These metrics are based on true positives (TP), false positives (FP), and false negatives (FN) and were calculated as follows: $$Precision=TP/(TP+FP)$$ 3 $$Recall=TP/(TP+FN)$$ 4 $$F1 score=2*\frac{Precision*Recall}{Precision+Recall}$$ 5 Quantitative assessment of the defoliation mapping was based on the 200 semi-random sample locations. Landsat, aerial, and Sentinel-2 imagery was used for visually classifying the sample points for 2015, 2016, and 2017, respectively. Accuracy assessment for the mortality modeling was based on the validation subsets (i.e. 20%) of the 2213 “ground-truth” tiles. The mortality models were further evaluated using Shapley Additive explanations (SHAP) plots and feature importance scores. SHAP plots indicate how the model uses a particular predictor and feature importance scores indicate the relative contribution of each predictor in the model classification. 3. Results 3.1. Defoliation During the 3-year Spongy moth outbreak (2015–2017), approximately 23% of Rhode Island's forest (312 km 2 ) experienced significant defoliation for one or more years, based on the 30 m resolution defoliation model. Repeated defoliation for a given area was uncommon. Of the forest land that experienced defoliation during the outbreak, 89.2% of forests (278 km 2 ) were defoliated only once. Only 10.6% (33 km 2 ) and 0.2% ( 75% forest cover, 33% (44,807 tiles) and 18% (24,050 tiles) of the tiles had defoliation over more than 25% and 75% of the tile area, respectively. Defoliation was distributed predominantly in the western part of Rhode Island, with repeated defoliation occurring sporadically, mostly in the southwestern and northwestern parts of the state. Detectable defoliation rarely occurred in coastal forests (Fig. 1 ). The defoliation model had good agreement with our visual interpretation of the aerial and satellite imagery for our 200 sample points. For accuracy assessment, we considered 1) complete defoliation only and 2) complete and partial defoliation. We had high confidence in our ability to visually detect complete defoliation in both satellite and aerial imagery, but lower confidence in our ability to detect partial defoliation in the relatively coarse Landsat and Sentinel-2 data. For complete defoliation, the model performed consistently throughout the outbreak period with F1 scores ranging from 0.84–0.86. When partial defoliation was also considered, the model accuracy was more variable with F1-scores ranging from 0.76–0.87; the best accuracy was associated with the 2016 data for which we were able to use aerial imagery to create more reliable validation data. Table 5 Accuracy assessment of defoliation model based on 200 sample locations. Year Validation Imagery Defoliation Type F1-score Sample Size 2015 Landsat Partial + complete 0.79 113 Complete 0.84 110 2016 Aerial Partial + complete 0.87 167 Complete 0.86 136 2017 Sentinel Partial + complete 0.76 188 Complete 0.86 154 3.2. Tree mortality Mortality was modeled for 1358 km 2 of forest land which corresponded to tiles with > 75% forest cover. Based on the models, tree mortality was highest in the western part of the state (Fig. 3 ). There was low tree mortality in the coastal areas and the eastern part of the state. For the 2-class model, 65% (883 km 2 ) and 35% (475 km 2 ) of forestland was classified as low and high mortality, respectively. For the 3-class model, 56% (760 km 2 ), 23% (312 km 2 ), and 21% (286 km 2 ) of forestland was classified as low, medium, and high mortality, respectively. Our best 2-class and 3-class models of tree mortality had overall accuracies of 82% and 65%, respectively (Table 6 , 7 ). When only defoliation was included as a predictive variable, accuracies were 72% and 56% for the 2-class and 3-class models, respectively. When the top 3 predictive variables were included, accuracies improved to 79% and 63% for the 2-class and 3-class models, respectively. For both 2- and 3-class models, precision tended to improve slightly with the inclusion of more predictive variables. However, the 21-variable models were not consistently better than the 7-variable model, and both models were only slightly better than the 3-variable model. For the 2-class model, recall tended to improve with the number of variables. However, for the 3-class models, there was no consistent pattern between recall and number of variables. Table 6 Accuracy assessment for 2-class models (low and high mortality) with varying numbers of variables. Metrics include overall accuracy (OA), cross-validation accuracy (XV), kappa (kp), F1 score (F1), precision (P), and recall (R). Number of variables OA XV kp Low High F1 P R F1 P R 21 0.81 0.78 0.61 0.77 0.82 0.73 0.84 0.81 0.87 7 0.82 0.77 0.63 0.77 0.78 0.76 0.86 0.85 0.86 3 0.79 0.77 0.56 0.73 0.77 0.70 0.83 0.80 0.85 1 0.72 0.74 0.42 0.67 0.66 0.67 0.75 0.76 0.75 Table 7 Accuracy assessment for 3-class models (low, medium, and high mortality) with varying numbers of variables. Metrics include overall accuracy (OA), cross-validation accuracy (XV), kappa (kp), F1 score (F1), precision (P), and recall (R). Number of variables OA XV kp Low Medium High F1 P R F1 P R F1 P R 21 0.64 0.60 0.46 0.67 0.63 0.72 0.47 0.54 0.41 0.75 0.71 0.80 7 0.65 0.61 0.47 0.65 0.67 0.64 0.50 0.45 0.51 0.77 0.77 0.70 3 0.63 0.60 0.43 0.65 0.66 0.64 0.48 0.45 0.51 0.74 0.77 0.48 1 0.56 0.55 0.33 0.60 0.51 0.73 0.30 0.23 0.40 0.68 0.67 0.70 The importance of the predictive variables was consistent among all the tree mortality models (Tables 8 , 9 , 10 ). The defoliation index was the most important predictor followed by coast proximity and canopy cover . For the models that included either all 21 variables or only the top 7 variables, the defoliation index was substantially more important than coast proximity , and coast proximity was substantially more important than canopy cover . In models with 7 or more variables, canopy cover was only slightly more important than evergreen cover . Evergreen cover, drought index , urban proximity , and lake proximity all had similar but relatively minor importance. Variables relating to soil and topography tended to be the least important predictors. Table 8 Relative importance of predictors used for the 21-variable models . Higher scores indicate greater importance and all scores sum to 1. Predictors 2-class Model 3-class Model Defoliation index 0.25 0.21 Coast proximity 0.16 0.13 Canopy cover 0.08 0.09 Evergreen cover 0.06 0.06 Drought index 0.06 0.06 Urban proximity 0.06 0.06 Lake proximity 0.05 0.05 Valley bottoms 0.04 0.04 Hilltops 0.04 0.04 Steep slopes 0.03 0.04 Hydric soils 0.02 0.03 Stony soils 0.02 0.03 Restrictive soils 0.02 0.02 Shallow bedrock 0.02 0.02 Steep south-facing slopes 0.01 0.02 Plantations 0.01 0.01 Till < 0.01 < 0.01 Outwash < 0.01 < 0.01 Over drained soils < 0.01 < 0.01 Eroded soils < 0.01 < 0.01 Ruderal forest < 0.01 < 0.01 Table 9 Relative importance of predictors used for the 7-variable models . Higher scores indicate greater importance, and all scores sum to 1. Predictors 2-class Model 3-class Model Defoliation index 0.30 0.27 Coast proximity 0.23 0.20 Canopy cover 0.11 0.13 Drought index 0.1 0.09 Evergreen cover 0.09 0.09 Urban proximity 0.09 0.11 Lake proximity 0.08 0.1 Table 10 Relative importance of predictors used for the 3-variable models . Higher scores indicate greater importance, and all scores sum to 1. Predictors 2-class Model 3-class Model Defoliation index 0.38 0.37 Coast proximity 0.36 0.36 Canopy cover 0.27 0.28 The SHAP analysis shows how individual variables influence the model results for each predicted class. For both the 2- and 3-class models, tiles that were classified as “low mortality” tended to have lower defoliation index , closer coast proximity , higher canopy cover , lower drought index , closer urban proximity , and higher evergreen cover . Conversely, tiles that were classified as “high mortality” tended to have higher defoliation index , further coast proximity , lower canopy cover , higher drought index , further urban proximity , and less evergreen cover . 4. Discussion This study used satellite-based defoliation mapping combined with geospatial environmental predictors to model tree mortality resulting from a Spongy moth outbreak in a mixed temperate deciduous forest. The models identified predictors that were associated with higher rates of mortality (e.g. defoliation index, coast proximity, canopy cover ). The inclusion of environmental predictors along with the defoliation index resulted in significant improvements to model performance. This finding is consistent with Campbell et al. (2020) found that GIS-based data explained an additional 17% of mortality beyond defoliation data alone for their semi-arid woodland study area. The performance of our models was on par with previous studies which focused on coniferous forests with relatively lower species diversity. Meddens et al. (2014), Goodwin et al. (2008), and Long et al. (2016) achieved accuracies of 70–80% for their coniferous forest study areas. Bergmüller et al. (2022) achieved an accuracy of 70% in Canadian mixed forests using very high-resolution imagery collected by unmanned aerial systems (UAS). Using GIS-based topographic data and field-based soil samples, Toledo et al. (2011) were able to explain 20% of tree mortality. Campbell et al. (2020) was one of the few studies that combined a defoliation metric with GIS-based environmental factors to model tree mortality. For their open-canopy semi-arid woodland, their model achieved an r 2 = 0.71. The SHAP analysis was generally consistent with our expectations of how various predictors should affect tree mortality. Higher tree mortality was associated with a higher defoliation index , further distance from coast, lower canopy cover, less evergreen cover, and closer proximity to urban cover. Since defoliation is the primary stressor for trees during a Spongy moth outbreak, the defoliation index was expected to be directly related to increased mortality. Evergreen cover was expected to be inversely related to mortality since coniferous species are not the preferred food source of Spongy moths. The higher mortality further from the coast may be due to higher temperatures and lower humidity found in inland areas (NCDC, 1971). Coastal proximity was much more important than drought index which suggests that other protective attributes (e.g. such as differing forest compositions) may be associated with coastal areas. The much higher resolution of the coastal proximity metric may also help explain its greater importance than the drought index. The higher mortality associated with lower canopy cover may be due to the increased solar heating of the ground and evaporation of soil moisture in more open forests. In our study area, forests with lower canopy cover could also signify previous disturbances that contributed to overall poor forest health. Our finding regarding canopy cover is consistent with Campbell et al. (2020) who studied semi-arid piñon-juniper woodlands. The protective effect of closer proximity to urban areas was somewhat unexpected. However, this relatively minor effect may be due to the greater care and management (e.g. prompt removal of dead trees) provided to trees in more urban environments. We unexpectedly found that soil-based predictors had little importance in our models for predicting tree mortality. Adverse soil conditions constrain maximum tree heights, slow growth rates, and stress trees through limited water or nutrient availability. Campbell et al. (2020) found that surface organic matter had moderate importance for predicting tree mortality but found that 29 other soil variables had very little value. The lack of importance for soils data in our study may reflect a limitation of GIS soils datasets. The minimum mapping unit of our dataset was around 1 ha which could omit much of the soil variation that would be relevant to mortality of individual trees. In addition, unfavorable individual soil characteristics were relatively uncommon in our study area and the majority of the tiles in our training/validation dataset had zeroes or very low values for soil-based predictors. The lack of variation may have made it less likely for Random Forest to find useful partitions of these predictors associated with varying levels of tree mortality. Combining the soil characteristics into a single metric may yield a more useful predictor for Random Forest. We found that topographic characteristics also had little importance in our models of tree mortality. Steeper south-facing slopes tend to receive more solar heating than other slope orientations which results in warmer and drier conditions that are likely to stress trees. However, the topography in our study area was relatively moderate with little area covered by steep south-facing slopes. The relatively infrequent occurrence of steep slopes in the study area may have made the predictor unlikely to be used effectively by Random Forest . However, slope and orientation may be more important factors in areas with rugged topography. Our models were only slightly improved when we included more than the 3 top predictors. Models with the 7 top predictors performed very similarly to models with the full set of predictors. The ability to use fewer predictors without sacrificing model performance is advantageous because it simplifies model development and improves efficiency. The fraction of a tree crown that is defoliated is likely to be an important factor in tree mortality. However, our defoliation metric was based primarily on complete defoliation (i.e. ~100% crown defoliation). Pasquarella et al. (2018) did map defoliation, from a Spongy moth outbreak, with differing levels of severity using a Tasseled Cap transform approach. However, we did not attempt to map differing severity levels because we found it difficult to assess partial defoliation visually based on the 10–30 m resolutions of the satellite imagery that we used for accuracy assessments. We also found significant annual variation in our baseline (i.e. pre-outbreak) NDVI values; thus, we chose a conservative threshold for defoliation to minimize commission error. This study was applied to a Spongy moth pest outbreak over a limited geographic area. Thus, the relevant predictors of mortality may be somewhat different for other forest pests or disease outbreaks in different areas. However, we did confirm the benefit of incorporating GIS-based predictors along with satellite-based mapping of defoliation in forest mortality models in temperate deciduous forests. The Random Forest tool was effective for using certain types of predictors in ways that are consistent with expectations. However, it seemed ineffective for using predictors that are relatively uncommon but still likely to be relevant (e.g. soils). Limitations in the spatial resolution of GIS data may also preclude the inclusion of important predictors in mortality models. Future work should explore whether uncommon, but likely relevant, features can be used more effectively in mortality models. 5. Conclusion Our study used satellite-based defoliation mapping and geospatial environmental data with Random Forest to model tree mortality from a Spongy moth outbreak in Rhode Island’s temperate deciduous forest. The best models achieved accuracies of 82% and 65% when predicting 2 classes (low/high) or 3 classes (low/med/high) of mortality, respectively. The inclusion of geospatial data improved model predictions by 7–10% compared to models based only on defoliation. The most important predictors in the models were the defoliation index , coastal proximity , and canopy cover . Models improved only slightly with the inclusion of more than 3 top predictors. Soil characteristics had very little contribution to the models which may be due to the coarse resolution (i.e. minimum mapping unit of 1 ha) and the tendency for individual adverse soil characteristics to be relatively uncommon. Topographic factors also had minimal influence on models which may be due to the relatively moderate topography of the study area. Although relevant predictors of tree mortality may vary somewhat for different regions and pest species, this study showed the benefit of Random Forest modeling that combines satellite-based monitoring with geospatial environmental data. Declarations Author Contribution Jason Parent and Liubov Dumarevskaya jointly conceived and designed the study. Liubov Dumarevskaya conducted data collection and analysis under Jason Parent's supervision. Liubov Dumarevskaya drafted the manuscript, and Jason Parent provided critical revisions for important intellectual content. Both authors approved the final version of the manuscript for submission. Acknowledgement The authors thank Rockwell Richards and Molly Ahern for their contributions in creating the training/validation datasets used in this study. References Anderegg, W. R. L., Hicke, J. A., Fisher, R. A., Allen, C. D., Aukema, J., Bentz, B., Hood, S., Lichstein, J. W., Macalady, A. K., McDowell, N., Pan, Y., Raffa, K., Sala, A., Shaw, J. D., Stephenson, N. L., Tague, C., Zeppel, M. 2015. Tree mortality from drought, insects, and their interactions in a changing climate. New Phytologist, 208(3), 674–683. DOI: https://doi.org/10.1111/nph.13477 Baker, W. L. 1941. Effect of Gypsy Moth Defoliation on Certain Forest Trees. Journal of Forestry, Volume 39, Issue 12, Pages 1017–1022. DOI: https://doi.org/10.1093/jof/39.12.1017 . Barbosa, P., Capinera, J.L., 1978. Population quality, dispersal and numerical change in the gypsy moth, Lymantria dispar (L.). Oecologia, 36, pp.203–209. DOI: 10.1007/BF00349809 Bergmüller, K. O., Vanderwel, M. C. 2022. Predicting Tree Mortality Using Spectral Indices Derived from Multispectral UAV Imagery. Remote Sensing, 14(9), 2195. DOI: https://doi.org/10.3390/rs14092195 . Bigsby, K.M., Ambrose, M.J., Tobin, P.C. Sills, E.O., 2014. The cost of gypsy moth sex in the city. Urban Forestry & Urban Greening, 13(3), pp.459–468. DOI: https://doi.org/10.1016/j.ufug.2014.05.003 Bottero, A., D'Amato, A.W., Palik, B.J., Bradford, J.B., Fraver, S., Battaglia, M.A., Asherin, L.A., 2017. Density-dependent vulnerability of forest ecosystems to drought. Journal of Applied Ecology, 54(6), pp.1605–1614. DOI: https://doi.org/10.1111/1365-2664.12847 Campbell, R. W. 1979. Gypsy Moth Influence on Forest. Agriculture Information Bulletin No. 423. United States Department of Agriculture, Forest Service, Pacific Northwest Forest and Range Experiment Station. Campbell, M. J., Dennison, P. E., Tune, J. W., Kannenberg, S. A., Kerr, K. L., Codding, B. F., Anderegg, W. R. L. 2020. A multi-sensor, multi-scale approach to mapping tree mortality in woodland ecosystems. Remote Sensing of Environment, Volume 245, 111853. DOI: https://doi.org/10.1016/j.rse.2020.111853 . Choat, B., Brodribb, T.J., Brodersen, C.R., Duursma, R.A., López, R. and Medlyn, B.E., 2018. Triggers of tree mortality under drought. Nature, 558(7711), pp.531–539. DOI: https://doi.org/10.1038/s41586-018-0240-x . National Oceanographic and Atmospheric Administration (NOAA). 2020. Coastal Change Analysis Program (C-CAP). High-Resolution Land Cover dataset for Rhode Island. Retrieved from [ https://chs.coast.noaa.gov/htdata/raster1/landcover/bulkdownload/hires/ri/] on July 2022. Das, A., Battles, J., Stephenson, N. L., van Mantgem, P. J. 2011. The contribution of competition to tree mortality in old-growth coniferous forests. Forest Ecology and Management, 261(7), 1203–1213. Davidson, C. B., Gottschalk, K. W., & Johnson, J. E. 1999. Tree mortality following defoliation by the European gypsy moth (Lymantria dispar L.) in the United States: a review. Forest science, 45(1), 74–84. Dorman, M., Perevolotsky, A., Sarris, D., Svoray, T. 2015. The effect of rainfall and competition intensity on forest response to drought: lessons learned from a dry extreme. Oecologia, 177, 1025–1038. DOI: 10.1007/s00442-015-3229-2 Dymond, J.R., 2010. Soil erosion in New Zealand is a net sink of CO 2 . Earth Surface Processes and Landforms, 35(15), pp.1763–1772. DOI: https://doi.org/10.1002/esp.2014 Dunn, C.P., Stearns, F., 1987. A comparison of vegetation and soils in floodplain and basin forested wetlands of southeastern Wisconsin. American Midland Naturalist, pp.375–384. Goodwin, N. R., Coops, N. C., Wulder, M. A., Gillanders, S., Schroeder, T. A., Nelson, T. 2008. Estimation of insect infestation dynamics using a temporal sequence of Landsat data. Remote Sensing of Environment, Volume 112, Issue 9, Pages 3680–3689. DOI: https://doi.org/10.1016/j.rse.2008.05.005 . Gottschalk, K. W., Colbert, J. J., & Feicht, D. L. 2007. Tree mortality risk of oak due to gypsy moth. European Journal of Forest Pathology. DOI: https://doi.org/10.1111/j.1439-0329.1998.tb01173.x . Guarín, A., & Taylor, A. H. 2005. Drought triggered tree mortality in mixed conifer forests in Yosemite National Park, California, USA. Forest Ecology and Management, 218(1–3), 229–244. DOI: https://doi.org/10.1016/j.foreco.2005.07.014 . Guggenmoos, S. 2003. Effects of tree mortality on power line security. Journal of Arboriculture, 29(4), 181–196. DOI: https://doi.org/10.48044/jauf.2003.022 Gunst, K., Weisberg, P.G., Yang, J., Fan. Y. Do denser forests have greater risk of tree mortality: A remote sensing analysis of density-dependent forest mortality, Forest Ecology and Management, Volume 359, 2016, Pages 19–32, ISSN 0378–1127, DOI: https://doi.org/10.1016/j.foreco.2015.09.032 . Enser, R., Gregg, D., Sparks, C., August, P., Jordan, P., Coit, J., Raithel, C., Tefft, B., Payton, B., Brown, C., LaBash, C., Comings, S., & Ruddock, K. 2011. Rhode Island Ecological Communities Classification. Technical Report. Rhode Island Natural History Survey, Kingston, RI. 33 pp. Eräjää, S., Halme, P., Kotiaho, J.S., Markkanen, A., Toivanen, T., 2010. The volume and composition of dead wood on traditional and forest fuel harvested clear-cuts. Silva Fennica, 44(2). DOI: 10.14214/sf.150 Fei, S., Morin, R. S., Oswalt, C. M., & Liebhold, A. M. 2019. Biomass losses resulting from insect and disease invasions in US forests. Proceedings of the National Academy of Sciences, 116(35), 17371–17376. Published on August 12, 2019. DOI: https://doi.org/10.1073/pnas.1820601116 . Franklin, J.F., Shugart, H.H. and Harmon, M.E., 1987. Tree death as an ecological process. BioScience, 37(8), pp.550–556. DOI: 10.2307/1310665 Horsley, S. B., Long, R. P., Bailey, S. W., Hallett, R. A., & Wargo, P. M. Health of eastern North American sugar maple forests and factors affecting decline. Northern Journal of Applied Forestry 19.1. 2002: 34–44. DOI: https://doi.org/10.1093/njaf/19.1.34 Hudgins, E.J., Liebhold, A.M. and Leung, B., 2017. Predicting the spread of all invasive forest pests in the United States. Ecology letters, 20(4), pp.426–435. DOI: https://doi.org/10.1111/ele.12741 . Ivantsova, E.D., Pyzhev, A.I. and Zander, E.V., 2019. Economic consequences of insect pests outbreaks in boreal forests: A literature review. Journal of Sibiric Federal University, 12(4), pp.627–642. DOI: 10.17516/1997-1370-0417 Kayama, M., Quoreshi, A.M., Kitaoka, S., Kitahashi, Y., Sakamoto, Y., Maruyama, Y., Kitao, M. and Koike, T., 2003. Effects of deicing salt on the vitality and health of two spruce species, Picea abies Karst., and Picea glehnii Masters planted along roadsides in northern Japan. Environmental pollution, 124(1), pp.127–137. DOI: DOI: 10.1016/s0269-7491(02)00415-3 . Kinahan, I.G., Grandstaff, G., Russell, A., Rigsby, C.M., Casagrande, R.A. and Preisser, E.L., 2020. A four-year, seven-state reforestation trial with eastern hemlocks ( Tsuga canadensis ) resistant to hemlock woolly adelgid ( Adelges tsugae ). Forests, 11(3), p.312. DOI: https://doi.org/10.3390/f11030312 United States Geological Survey (USGS). Available at https://earthexplorer.usgs.gov/ (last accessed on November 2022). Long, J. A., Lawrence, R. L. 2016. Mapping Percent Tree Mortality Due to Mountain Pine Beetle Damage. Forest Science, Volume 62, Issue 4, Pages 392–402. DOI: https://doi.org/10.5849/forsci.15-046 . Ma, Q., Su, Y., & Guo, Q. 2017. Comparison of Canopy Cover Estimations From Airborne LiDAR, Aerial Imagery, and Satellite Imagery. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 10(9), 4225–4236. DOI: https://doi.org/10.1109/JSTARS.2017.2711482 . Meddens, A. J. H., Hicke, J. A., Vierling, L. A., Hudak, A. T. 2013. Evaluating methods to detect bark beetle-caused tree mortality using single-date and multi-date Landsat imagery. Remote Sensing of Environment, Volume 132, Pages 49–58. DOI: https://doi.org/10.1016/j.rse.2013.01.002 . Meddens, A. J. H., Hicke, J. A. 2014. Spatial and temporal patterns of Landsat-based detection of tree mortality caused by a mountain pine beetle outbreak in Colorado, USA. Forest Ecology and Management, Volume 322, Pages 78–88. DOI: https://doi.org/10.1016/j.foreco.2014.02.037 . Meng, R., Gao, R., Zhao, F., Huang, C., Sun, R., Lv, Z., Huang, Z. 2022. Landsat-based monitoring of southern pine beetle infestation severity and severity change in a temperate mixed forest. Remote Sensing of Environment, 269, 112847. DOI: https://doi.org/10.1016/j.rse.2021.112847 . Murray, B.C., Pendleton, L., Jenkins, W.A. and Sifleet, S., 2011. Green payments for blue carbon: economic incentives for protecting threatened coastal habitats. Nicholas Institute for Environmental Policy Solutions, Duke University, Durham. Navarro-Cerrillo, R.M., Varo-Martínez, M.Á., Acosta, C., Rodriguez, G.P., Sánchez-Cuesta, R., Ruiz Gómez, F.J. 2019. Integration of WorldView-2 and airborne laser scanning data to classify defoliation levels in Quercus ilex L. Dehesas affected by root rot mortality: Management implications. Forest Ecology and Management, 451, 117564. DOI: https://doi.org/10.1016/j.foreco.2019.117564 . National Climatic Data Center (NCDC). 1971. Climates of the States, Volume 1. Distributed by the U.S. Govt. Print. Off., Washington. (Climatography of the United States). Retrieved from https://books.google.com/books?id=pfHFwgEACAAJ . National Drought Mitigation Center. (2015–2017). U.S. Drought Monitor (USDM): Weekly ratings of drought across the U.S for the summer period. Øystein H. Opedal, W. Scott Armbruster & Bente J. Graae. 2015. Linking small-scale topography with microclimate, plant species diversity and intra-specific trait variation in an alpine landscape, Plant Ecology & Diversity, 8:3, 305–315, DOI: 10.1080/17550874.2014.987330 Parotta, J., Knowles, O., Wunderle, J.M., 1997. Floristic diversity development in a 10-year-old restoration forest on a bauxite mined site in Amazonia. Forest Ecol. Manage, 99, pp.21–42. DOI: https://doi.org/10.1016/S0378-1127(97)00192-8 Pasquarella, V.J., Elkinton, J.S. & Bradley, B.A. Extensive gypsy moth defoliation in Southern New England characterized using Landsat satellite observations. Biol Invasions 20, 3047–3053. 2018. DOI: https://doi.org/10.1007/s10530-018-1778-0 . Percy, K.E., Ferretti, M., 2004. Air pollution and forest health: toward new monitoring concepts. Environmental pollution, 130(1), pp.113–126. Poulos, H.M. and Camp, A.E., 2010. Decision support for mitigating the risk of tree induced transmission line failure in utility rights-of-way. Environmental management, 45, pp.217–226. DOI: https://doi.org/10.1007/s00267-009-9422-5 . Quirion, B.R., Domke, G.M., Walters, B.F., Lovett, G.M., Fargione, J.E., Greenwood, L., Serbesoff-King, K., Randall, J.M. and Fei, S., 2021. Insect and disease disturbances correlate with reduced carbon sequestration in forests of the contiguous United States. Frontiers in Forests and Global Change, 4, p.716582. DOI: https://doi.org/10.3389/ffgc.2021.716582 . Ramsfield, T.D., Bentz, B.J., Faccoli, M., Jactel, H. and Brockerhoff, E.G., 2016. Forest health in a changing world: effects of globalization and climate change on forest insect and pathogen impacts. Forestry, 89(3), pp.245–252. DOI: https://doi.org/10.1093/forestry/cpw018 . De Reu, J., Bourgeois, J., Bats, M., Zwertvaegher, A., Gelorini, V., De Smedt, P., Chu, W., Antrop, M., De Maeyer, P., Finke, P., Van Meirvenne, M., Verniers, J., Crombé, P. 2013. Application of the topographic position index to heterogeneous landscapes. Catena, 103, 31–39. DOI: https://doi.org/10.1016/j.catena.2012.11.008 . Rhode Island Geographic Information Systems (RIGIS). 2023. Rhode Island Maps and Data Geospatial Hub. Available at www.rigis.org (last accessed on November 2023). Rullan-Silva, C., Olthoff, A., Delgado de la Mata, J., & Pajares-Alonso, J. (2013). Remote Monitoring of Forest Insect Defoliation - A Review -. Forest Systems, 22(3), 377–391. DOI: https://doi.org/10.5424/fs/2013223-04417 . Seidl, R., Schelhaas, M.J., Rammer, W. and Verkerk, P.J., 2014. Increasing forest disturbances in Europe and their impact on carbon storage. Nature climate change, 4(9), pp.806–810. DOI: https://doi.org/10.1038/nclimate2318 . Shearman, T. M., Varner, J. M., Hood, S. M., Cansler, C. A., Hiers, J. K. 2019. Modelling post-fire tree mortality: Can random forest improve discrimination of imbalanced data? Ecological Modelling, Volume 414, 108855. DOI: https://doi.org/10.1016/j.ecolmodel.2019.108855 . Scikit-Learn: Machine Learning in Python. Random Forests in Python. Retrieved from https://scikit-learn.org/stable/modules/generated/sklearn.ensemble.RandomForestClassifier.html on November 2022. Stravinskienė, V., Bartkevičius, E., Abraitienė, J., Dautartė, A. 2018. Assessment of Pinus sylvestris L. tree health in urban forests at highway sides in Lithuania. Global Ecology and Conservation, 16, e00517. DOI: https://doi.org/10.1016/j.gecco.2018.e00517 . de Toledo, J. J., Magnusson, W. E., Castilho, C. V., & Nascimento, H. E. M. 2012. Tree mode of death in Central Amazonia: Effects of soil and topography on tree mortality associated with storm disturbances. Forest Ecology and Management, Volume 263, Pages 253–261. DOI: https://doi.org/10.1016/j.foreco.2011.09.017 . Trumbore, S., Brando, P., Hartmann, H., 2015. Forest health and global change. Science, 349(6250), pp.814–818. DOI: 10.1126/science.aac6759 . Verbesselt, J., Robinson, A., Stone, C. and Culvenor, D., 2009. Forecasting tree mortality using change metrics derived from MODIS satellite data. Forest Ecology and Management, 258(7), pp.1166–1173. DOI: https://doi.org/10.1016/j.foreco.2009.06.011 . Zhan, Z., Yu, L., Li, Z., Ren, L., Gao, B., Wang, L., Luo, Y. 2020. Combining GF-2 and Sentinel-2 Images to Detect Tree Mortality Caused by Red Turpentine Beetle during the Early Outbreak Stage in North China. Forests, 11(2), 172. DOI: https://doi.org/10.3390/f11020172 . Zhang, X., Jiao, J. J., & Guo, W. 2022. How Does Topography Control Topography-Driven Groundwater Flow? Geophysical Research Letters, 49(20), e2022GL101005. DOI: https://doi.org/10.1029/2022GL101005 United States Census Bureau (USCB). 2020. Historical Population Density Data (1910–2020). Retrieved from https://www.census.gov/data/tables/time-series/dec/density-data-text.html on November 2022. Additional Declarations No competing interests reported. Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-4378454","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":299647126,"identity":"534d668f-fa59-4732-af09-bbe2b93c576b","order_by":0,"name":"Liubov Dumarevskaya","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA6klEQVRIiWNgGAWjYBACPoYEKIsHRFQwMLABKQl8WthQtZwhWQtjG4SNXwt7jumGn3tsEuf3HH4m8XPe4Xw+BuaDt3nwaeF5Y3az51la4oazbWaSvdsOW7YxsCVb49UikWN2g+fAYWMDfgYzCd5thw3YGHjMpAlpufnnwH9j+X72b5J/54C08H8jqOU2z4EDcgxne8ykeRvAtrDh18LzrOy2zIFkOYMzZ4qtZY6lG7AxsxlbzsGjhZ89edvNNwfseOR70jfefFNjbSDf3vzwxhs8WpABCyQ6mIlUDlb7gQTFo2AUjIJRMIIAAGCaRWUOl6JAAAAAAElFTkSuQmCC","orcid":"","institution":"University of Rhode Island","correspondingAuthor":true,"prefix":"","firstName":"Liubov","middleName":"","lastName":"Dumarevskaya","suffix":""},{"id":299647128,"identity":"84e51f85-d9ab-43e1-85f0-1832c3385b1a","order_by":1,"name":"Jason R. Parent","email":"","orcid":"","institution":"University of Rhode Island","correspondingAuthor":false,"prefix":"","firstName":"Jason","middleName":"R.","lastName":"Parent","suffix":""}],"badges":[],"createdAt":"2024-05-06 17:27:38","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-4378454/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-4378454/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":56482207,"identity":"4e245e87-2805-426a-9837-f202cff7a25d","added_by":"auto","created_at":"2024-05-14 18:58:23","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":849804,"visible":true,"origin":"","legend":"\u003cp\u003eOur study area was the state of Rhode Island\u003cstrong\u003e.\u003c/strong\u003e\u003c/p\u003e","description":"","filename":"floatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-4378454/v1/2d6eb6590bc323564e4711b6.png"},{"id":56482208,"identity":"e96a69dd-d23f-4d16-9309-fea2ff77a49b","added_by":"auto","created_at":"2024-05-14 18:58:23","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":251746,"visible":true,"origin":"","legend":"\u003cp\u003eDefoliation in Rhode Island after Spongy moth invasion in 2015-2017.\u003c/p\u003e","description":"","filename":"floatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-4378454/v1/5c01772ad739f07951524bcf.png"},{"id":56482210,"identity":"80b6fc30-3479-4c89-8e76-b4375b1f5fe1","added_by":"auto","created_at":"2024-05-14 18:58:23","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":728815,"visible":true,"origin":"","legend":"\u003cp\u003eMaps of mortality prediction based on 2-class model (left) and 3-class model (right).\u003c/p\u003e","description":"","filename":"floatimage3.png","url":"https://assets-eu.researchsquare.com/files/rs-4378454/v1/5dab7a47ac3340717ba006a4.png"},{"id":56482209,"identity":"576073cd-7746-404e-936a-cc674100af0b","added_by":"auto","created_at":"2024-05-14 18:58:23","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":438542,"visible":true,"origin":"","legend":"\u003cp\u003eSHAP analysis of the 7 most important predictors for the low-mortality (top) and high mortality (bottom) classes of the \u003cstrong\u003e2-class model\u003c/strong\u003e. Positive SHAP values (x-axis) indicate the feature values (high - red, low - blue) that are more strongly associated with the class.\u003c/p\u003e","description":"","filename":"floatimage4.png","url":"https://assets-eu.researchsquare.com/files/rs-4378454/v1/844541be4a8fd0e1460f4e2c.png"},{"id":56482211,"identity":"62ee077f-c5a5-498c-9406-35132283c200","added_by":"auto","created_at":"2024-05-14 18:58:23","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":428704,"visible":true,"origin":"","legend":"\u003cp\u003eSHAP analysis of the 7 most important predictors for the low-mortality (top) and high mortality (bottom) classes of the \u003cstrong\u003e3-class model\u003c/strong\u003e. Positive SHAP values (x-axis) indicate the feature values (high - red, low - blue) that are more strongly associated with the class.\u003c/p\u003e","description":"","filename":"floatimage5.png","url":"https://assets-eu.researchsquare.com/files/rs-4378454/v1/3ec7eb015816605b596c7a94.png"},{"id":56792948,"identity":"8883f847-9fa5-436f-9c2c-539921c3f921","added_by":"auto","created_at":"2024-05-20 14:08:23","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":3546585,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-4378454/v1/fdc3bc21-8d6d-4167-a108-993e55fc7a7a.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"Modeling Spongy Moth Forest Mortality in Rhode Island Temperate Deciduous Forest","fulltext":[{"header":"1. Introduction","content":"\u003cp\u003eInvasive pests have far-reaching ecological and economic impacts, affecting global forest communities and carbon storage (Seidl et al., 2014). Globally, Quirion et al. (2021) found that forests affected by pest invasion sequester 69% less carbon, on average, compared to unaffected forests. In the U.S. alone, forest pest activity results in annual biomass losses of 5.5 TgC (Fei et al., 2019), which is roughly equivalent to 11,800 hectares of temperate forest land (based on Murray et al., 2011). One important forest pest of the U.S. is Spongy moth (\u003cem\u003eLymantria dispar dispar\u003c/em\u003e) which can cause severe tree defoliation that leads to mortality. From 1994\u0026ndash;2010, Spongy moth caused an estimated 898 TgC of biomass loss in the U.S. (Fei et al., 2019), which is equivalent to roughly 1.9 M hectares of temperate forest (Murray et al., 2011). In a single U.S. city (i.e. Baltimore, MD), the damages caused by Spongy moth was estimated to range from \u003cspan\u003e$\u003c/span\u003e5.5\u0026ndash;63.7 M per year of outbreak, depending on environmental and management scenarios (Bigsby et al., 2014).\u003c/p\u003e \u003cp\u003eSatellite and aerial imagery allow for a regular monitoring of forest health and tree mortality (Verbesselt et al., 2009). Sentinel-2 and Landsat satellites provide moderate-resolution imagery over large areas. Summertime aerial imagery is acquired regularly by various agencies including the USDA\u0026rsquo;s National Agriculture Inventory Program (NAIP). Thus, remote sensing provides an important tool for monitoring forest health which can allow foresters to mitigate the impact of insect invasions through forest management (Hudgins et al., 2017). Mitigation can include direct prevention strategies such as biological controls and insecticides; indirect controls such as forest thinning, prescribed fire, and quarantines (Ivantsova et al., 2019); or replanting with pest-resistant varieties of the affected species (Kinahan et al., 2020). Tree mortality maps can guide tree removal programs to protect public safety and infrastructure, maintain aesthetics (Poulos et al., 2010; Guggenmoose et al., 2003), improve forest regeneration (Parotta et al., 1997), or convert forests with high mortality to early successional wildlife habitat.\u003c/p\u003e \u003cp\u003eDefoliation by Spongy moth and other leaf-eating insects, has a strong impact on tree mortality (Gottschalk et al., 2007; Baker et al., 1941). Davidson et al. (1999) reported that Spongy moth invasion caused a die-off of up to 47% of deciduous trees in outbreak areas in the Eastern U.S. When more than 75% of a tree crown is defoliated, a tree may produce a second set of leaves in the same growing season which depletes energy reserves and decreases chances of surviving the invasion (Davidson et al., 1999), especially if the tree is defoliated multiple times (Campbell, 1979). Other pests, such as borers, are attracted to stressed trees which further increases chances of tree mortality (Davidson 1999).\u003c/p\u003e \u003cp\u003eResearchers have mapped defoliation, based on satellite and aerial imagery, and used these maps to model tree mortality (Meddens et al., 2014; Goodwin et al., 2008; Long et al., 2016; Bergm\u0026uuml;ller et al., 2022; Meng et al., 2022; Zhan et al., 2020). In the Rocky Mountains, Meddens et al. (2014) mapped dead pine trees with Landsat-based vegetation indexes and linear regression models (r\u003csup\u003e2\u003c/sup\u003e\u0026thinsp;=\u0026thinsp;0.77). In boreal forests, Goodwin et al. (2008) classified dead pine trees, with 75% accuracy, using Normalized Difference Moisture Index derived from Landsat imagery. In spruce-dominated forests, Long et al. (2016) classified dead trees, with 90% accuracy, using a combination of high-resolution aerial imagery and moderate-resolution satellite imagery. In the forests of western Canada, Bergmuller et al. (2022) used multispectral drone imagery to map tree mortality with approximately 70% and 80% accuracy for deciduous and coniferous trees, respectively. In China, Meng et al. (2022) used multispectral satellite imagery to map Southern Pine Beetle infestations with an accuracy\u0026thinsp;\u0026gt;\u0026thinsp;80%. In pine forests in northern China, Zhan et al. (2020) used GF-2 and Sentinel-2 imagery to map tree mortality with 78% accuracy. Pasquarella et al. (2018) mapped defoliation from a Spongy moth outbreak with Landsat imagery using a tasseled-cap greenness index, but they did not model actual mortality.\u003c/p\u003e \u003cp\u003eDefoliation is a driving factor in tree mortality which can be compounded by various environmental stressors. Soil and geology characteristics may directly impact water availability, root depth, and nutrient availability, which can stress trees and decrease survival during defoliation events. Topographic position (i.e. hilltop, mid-slope or valley) can affect light and precipitation distribution, ground surface temperature, soil conditions, and depth of the groundwater which collectively determine water availability (Zhang et al., 2022; \u0026Oslash;ystein et al, 2015; Dunn et al., 1987). When analyzing factors related to the cause of death of individual trees in tropical moist forests of northern Amazonia, Toledo (2011) found that soil characteristics and topographic relief together accounted for 20% of the mortality variation. They also found that tree mortality was higher on steep slopes and on sandy soils in valleys while mortality was lower on plateaus with clay soils. Drought conditions reduce water availability and adds stress to defoliated tree (Bottero et al., 2017, Ramsfield et al., 2016; Choat et al., 2018). Guarin et al. (2005) found correlations between multiple years of drought and tree mortality. Tree in close proximity to roads and urban areas may experience greater stress and mortality due to exposure to de-icing salt (Kayama et al., 2003, Horsley et al., 2002) and air pollution (Trumbore et al., 2015, Percy et al., 2004). Forests near urban areas may have higher exposure to invasive plants or pests (Stravinskienė et al., 2018). In New England (U.S.), Davidson et al. (1999) reported that overall tree health significantly affected survival from Spongy moth outbreaks with mortality rates of 36% and 7% for unhealthy and healthy trees, respectively.\u003c/p\u003e \u003cp\u003eForest characteristics can influence tree mortality during pest outbreaks. Some invasive pests have species-specific feeding preferences and, thus, preferred species tend to be more heavily affected (Davidson et al., 1999; Barbosa et al., 1978). Canopy cover has been found to be an important factor in mortality prediction, although the effect was not consistent across different forest types and environments (Campbell et al., 2020, Gunst et al., 2016, Dorman et al., 2015, Das et al., 2011). In woodland areas of Utah, Campbell et al. (2020) found that high mortality is more likely to occur in areas with low-to-moderate canopy cover (5\u0026ndash;25%). Gunst et al. (2016) found that higher canopy cover was positively associated with tree mortality in xeric pine forests but negatively associated with tree mortality in red fir forests, during wet years. Das et al. (2011) found that canopy cover did not have a major effect on mortality in forests in the western U.S. where mortality was driven by insects and diseases. Dorman et al. (2015) found no association between canopy cover and mortality in pine forests.\u003c/p\u003e \u003cp\u003eA few studies have included both tree mortality and environmental geospatial data as predictors in tree mortality models. In open canopy semi-arid woodlands, Campbell et al. (2020) found that the inclusion of environmental factors significantly improved predictions of tree mortality when combined with defoliation mapped from high-resolution aerial imagery and Landsat imagery. Navarro-Cerrillo et al. (2019) mapped dead holm oak trees (\u003cem\u003eQuercus ilex\u003c/em\u003e) with 87% accuracy using WorldView-2 satellite imagery and light detection and ranging (LiDAR) data. They identified a link between high mortality and specific soil properties in Spanish woodlands impacted by root rot decline.\u003c/p\u003e \u003cp\u003eMost tree mortality models have been based solely on defoliation metrics and they tended to be conducted in coniferous forests which were dominated by few tree species (Meddens et al., 2013; Meddens et al., 2014; Bergm\u0026uuml;ller et al., 2022; Goodwin et al., 2008; Shearman et al., 2019). Studies that modeled tree mortality for mixed deciduous forests, with a wide variety of species, tended use very high-resolution drone imagery (Bergm\u0026uuml;ller et al., 2022, Long et al., 2016), which is not practical to collect for large study areas. Some studies have included both environmental factors and satellite-based defoliation maps in tree mortality models (Campbell et al. 2020, Navarro-Cerrillo et al., 2019); however, we are not aware of any studies that have applied these methods in temperate deciduous forests.\u003c/p\u003e \u003cp\u003eThe objective of this paper is to model tree mortality resulting from a 2015\u0026ndash;2017 Spongy moth outbreak in the temperate deciduous forest in Rhode Island, located in the northeastern U.S. We use a random forest approach to model mortality based on defoliation and environmental factors. We map defoliation from Landsat imagery and include geospatial data representing topography, climate, soil, and vegetation characteristics in modeling mortality. This research explores the importance of defoliation and environmental factors, represented by geospatial data, in predicting tree mortality from forest pest outbreaks.\u003c/p\u003e"},{"header":"2. Methods","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003e2.1. Study Area\u003c/h2\u003e \u003cp\u003eOur study area was the state of Rhode Island which has a total area of 3,144 km\u003csup\u003e2\u003c/sup\u003e. Rhode Island is largely covered by temperate deciduous forests dominated by oak (\u003cem\u003eQuercus sp.)\u003c/em\u003e and hickory (\u003cem\u003eCarya sp.)\u003c/em\u003e species (Enser et al., 2011). Other common species include white pine (\u003cem\u003ePinus strobus\u003c/em\u003e), sugar maple (\u003cem\u003eAcer saccharum\u003c/em\u003e), red maple (\u003cem\u003eAcer rubrum\u003c/em\u003e), American beech (\u003cem\u003eFagus grandifolia\u003c/em\u003e), hemlock (\u003cem\u003eTsuga canadensis\u003c/em\u003e), birch (\u003cem\u003eBetula sp\u003c/em\u003e.), elm (\u003cem\u003eUlmus americana\u003c/em\u003e), white ash (\u003cem\u003eFraxinus americana)\u003c/em\u003e, and linden tree \u003cem\u003e(Tilia americana)\u003c/em\u003e (Enser et al., 2011). Coniferous forests of Rhode Island are dominated by eastern white pine (\u003cem\u003ePinus strobus\u003c/em\u003e) and occur mostly in the southern part of state (Enser et al., 2011). The topography of Rhode Island mostly consists of gently rolling uplands with elevations up to 60 m and hillier upland terrain in the northwest with elevations of 60 to 200 m. The state includes humid continental, humid subtropical, and oceanic climate types (NCDC, 1971). Mean annual temperatures in Rhode Island range from 8-12\u003csup\u003eo\u003c/sup\u003eC and annual precipitation ranges from 100\u0026ndash;150 cm with lower precipitation amounts in coastal areas. Summers are characterized by periods (i.e. \u0026gt; 2 weeks) of little to no precipitation.\u003c/p\u003e \u003cp\u003eRhode Island is the second most densely populated state in the U.S. with an average of 393 persons / km\u003csup\u003e2\u003c/sup\u003e (USCB, 2020). The state contains 1784 km\u003csup\u003e2\u003c/sup\u003e of forest land which generates an estimated \u003cspan\u003e$\u003c/span\u003e408 M in forest products and \u003cspan\u003e$\u003c/span\u003e720 M annually in recreational value (RIDEM, 2019). In addition, the forests provide erosion and flood control and other ecosystem services (RIDEM, 2019).\u003c/p\u003e \u003cp\u003eThe study area experienced a severe spongy moth outbreak in 2015\u0026ndash;2017. This outbreak defoliated approximately 4386 km\u003csup\u003e2\u003c/sup\u003e of forest in New England which made it the largest outbreak in 30 years (Pasquarella et al., 2018). This outbreak was caused by unusually dry springs in 2014, 2015, and 2016. Those dry conditions suppressed a growth of fungus \u003cem\u003eEntomophaga maimaiga\u003c/em\u003e which typically controls spongy moth populations in the region (Pasquarella et al., 2018).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec4\" class=\"Section2\"\u003e \u003ch2\u003e2.2 Data and Software\u003c/h2\u003e \u003cp\u003eTo map defoliation, we used atmospherically corrected (level 2) Landsat 8 satellite imagery acquired in mid-summer (i.e. late June to early August) from 2009\u0026ndash;2017 (USGS, 2022). Landsat imagery has a 30 m spatial resolution for the visible and near-infrared wavelengths. Images corresponded as closely as possible to early July after spongy moth activity reached its peak and before significant refoliation could occur. We acquired cloud-free imagery for 2009, 2011, 2013, and 2016. No cloud-free images were available for the remaining years so we to used multiple images to create cloud-free composites. The composite images were created by using a cloud mask layer to identify cloud pixels in the primary images and replacing those pixels with cloud-free image pixels from the secondary images. Secondary images had acquisition dates within 4 weeks of the primary images.\u003c/p\u003e \u003cp\u003eWe used high-resolution summertime aerial imagery for July 2016 (RIGIS, 2023) to evaluate defoliation mapping for the corresponding year. The imagery had a spatial resolution of 7.6 cm and includes visible and near-infrared wavelengths. To evaluate defoliation mapping for 2015 and 2017, we used atmospherically corrected Sentinel-2 imagery (USGS, 2022). Sentinel-2 imagery has a spatial resolution of 10 m for the visible and near-infrared wavelengths. To create training and validation for the tree mortality models, we used aerial imagery acquired during in leaf-on summer period in July-August of 2019 with visible (RGB) bands and a 7.6 cm spatial resolution (RIGIS, 2023).\u003c/p\u003e \u003cp\u003eVarious datasets were used to create environmental metrics. Most of these datasets were downloaded from the Rhode Island Geographic Information System and included \u003cem\u003eSoils\u003c/em\u003e, \u003cem\u003eGlacial Deposits\u003c/em\u003e, \u003cem\u003eEcological Communities of Rhode Island\u003c/em\u003e, \u003cem\u003e2011 Rhode Island Statewide LiDAR, Freshwater Lakes, Ponds and Reservoirs\u003c/em\u003e, and \u003cem\u003eCoastal Waters\u003c/em\u003e (RIGIS, 2023). We also used the \u003cem\u003eU.S. Drought Monitor\u003c/em\u003e data (NDMC, 2022) and a 1 m resolution land cover from the National Oceanographic and Atmospheric Administration Coastal Change Analysis Program (CCAP) (NOAA, 2016). We used weekly Drought Monitor data for the summertime (June 1 \u0026ndash; Sept. 1) during the spongy moth outbreak (2015\u0026ndash;2017). The LiDAR data were acquired during leaf-off conditions in March and had a point density of 2\u0026thinsp;+\u0026thinsp;pts/m\u003csup\u003e2\u003c/sup\u003e. The ground-classified points in the LiDAR data were used to create a 1-m resolution digital elevation model (DEM). For all the GIS-based data processing and analysis we used ArcGIS Pro with Python 3.11.3. The Python modules SciKit-Learn, Seaborn, and SHAP were used to run \u003cem\u003eRandom Forest\u003c/em\u003e analysis, calculate related metrics, and create figures for tree mortality modelling.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec5\" class=\"Section2\"\u003e \u003ch2\u003e2.3. Predictors of Tree Mortality\u003c/h2\u003e \u003cp\u003eWe created a 100 m x 100 m (1 ha) polygon grid covering the study area to serve as the analysis units for modeling. Tiles were included in the analysis if they had at least 75% forest cover (i.e. deciduous, coniferous, mixed, wetland), based on the CCAP land cover data. Most of the environmental metrics represent the fraction of a given tile that corresponds to the particular feature-of-interest. Altogether, we developed 21 environmental metrics which characterized soil properties, topography, climate, forest characteristics, and proximity to resources and stressors (Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e).\u003c/p\u003e \u003cdiv id=\"Sec6\" class=\"Section3\"\u003e \u003ch2\u003e2.3.1. Defoliation\u003c/h2\u003e \u003cp\u003eWe used Normalized Difference Vegetation Index (NDVI) to map defoliation with 30 m spatial resolution based on Landsat imagery (Rullan-Silva et al., 2013). To establish a baseline NDVI for the pre-outbreak forest, we calculated NDVI for the 2009, 2011, and 2013 summer images. We estimated annual pre-outbreak variation in NDVI based on a sample of 200 points. The sample included 100 points in locations with obvious defoliation and 100 points in locations with no apparent defoliation, based on the 2016 summertime aerial imagery. Points were semi-randomly distributed to encompass the extent of forest land in Rhode Island. The NDVI values for the pixels corresponding to the sample point locations were extracted from the pre-outbreak imagery. We calculated the average pre-outbreak NDVI values for each pixel as well as the standard deviations for each year. For each outbreak year (2015\u0026ndash;2017), the NDVI rasters were subtracted from the pre-outbreak average NDVI raster. We considered NDVI differences of \u0026gt;\u0026thinsp;0.09 to correspond to significant defoliation. This threshold was double the largest pre-outbreak standard deviation which helped ensure that the differences exceeded the typical yearly variations.\u003c/p\u003e \u003cp\u003eWe calculated a defoliation index for the 100 m tiles which weighted multiple years of defoliation more heavily than a single year of defoliation. The defoliation index was calculated as:\u003cdiv id=\"Equ1\" class=\"Equation\"\u003e\u003cdiv format=\"TEX\" class=\"mathdisplay\" id=\"FileID_Equ1\" name=\"EquationSource\"\u003e\n$$Defoliation Index=fD1 + 2*fD2 + 3*fD3$$\u003c/div\u003e\u003cdiv class=\"EquationNumber\"\u003e1\u003c/div\u003e\u003c/div\u003e\u003c/p\u003e \u003cp\u003ewhere \u003cem\u003efD1\u003c/em\u003e, \u003cem\u003efD2\u003c/em\u003e, and \u003cem\u003efD3\u003c/em\u003e are the fraction of area within each tile that experienced defoliation for 1, 2, or 3 years respectively (Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec7\" class=\"Section3\"\u003e \u003ch2\u003e2.3.2. Environmental metrics\u003c/h2\u003e \u003cp\u003eCanopy cover can have positive or negative impacts on tree mortality depending on the ecosystem. We used the LiDAR point cloud to calculate canopy cover based on the first-return cover index (FRCI) (Ma et al., 2017). This metric was calculated, based on the LiDAR points within each tile, as the ratio of first returns that intercept the canopy (First\u003csub\u003eCanopy\u003c/sub\u003e) to the total number of first returns(First\u003csub\u003eTotal\u003c/sub\u003e):\u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\)\u003c/span\u003e\u003c/span\u003e\u003cdiv id=\"Equ2\" class=\"Equation\"\u003e\u003cdiv format=\"TEX\" class=\"mathdisplay\" id=\"FileID_Equ2\" name=\"EquationSource\"\u003e\n$$FRCI={\\Sigma }\\text{F}\\text{i}\\text{r}\\text{s}\\text{t}\\text{C}\\text{a}\\text{n}\\text{o}\\text{p}\\text{y} / {\\Sigma }\\text{F}\\text{i}\\text{r}\\text{s}\\text{t}\\text{T}\\text{o}\\text{t}\\text{a}\\text{l}$$\u003c/div\u003e\u003cdiv class=\"EquationNumber\"\u003e2\u003c/div\u003e\u003c/div\u003e\u003c/p\u003e \u003cp\u003eFirst returns indicate the location where the LiDAR pulse first intercepted an object. We considered first returns with heights\u0026thinsp;\u0026gt;\u0026thinsp;3 m to correspond to the canopy. FRCI values range from 0 (no cover) to 1 (complete cover) (Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eThe amount of coniferous tree cover can be important factor because Spongy moth prefers to feed on deciduous trees. Also, ruderal and plantation forests have different tolerances for stressors than later successional forests or forests that are not heavily managed. Thus, we calculated the fractions of each tile covered by evergreen, ruderal, and plantation forests, based the Ecological Communities dataset (Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eDrought conditions can have a major impact on tree survival when it coincides with other stressors such as defoliation. We calculated a \u003cem\u003edrought index\u003c/em\u003e during the peak growing season (i.e. June through August) for the 2015\u0026ndash;2017 outbreak period (Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e). The drought data consisted of polygons with attributes ranging from 0\u0026ndash;6 to indicate conditions ranging from no drought to severe drought (NDMC 2022). The polygons were intersected with the 100 m grid to join the weekly drought indices to our study tiles. The \u003cem\u003edrought index\u003c/em\u003e was the average of the weekly drought ratings over all the summer seasons of the outbreak period for each tile.\u003c/p\u003e \u003cp\u003eCoastal environments have differing climatic conditions compared to inland environments. To represent \u003cem\u003ecoast proximity\u003c/em\u003e, we calculated distances from the coastline as defined by the Rhode Island Coastal Waters dataset. The distances within each tile were averaged to create the metric (Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eUrban proximity can stress trees through exposure to air pollution, road salt, heat island effects, and invasive pests (Stravinskiene et al., 2018, Trumbore et al., 2015, Kayama et al., 2003, Percy et al., 2004, Horsley et al., 2002). We calculated \u003cem\u003eurban distance\u003c/em\u003e as distance from developed land cover in the CCAP land cover dataset. The distances within each tile were averaged to create the metric (Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eProximity to waterbodies can influence soil moisture levels, nutrient availability, and the prevalence of certain pests and diseases. We used the Freshwater Lakes, Ponds and Reservoirs dataset to calculate distance from waterbodies. The distances within each tile were averaged to create the \u003cem\u003eLake proximity\u003c/em\u003e metric (Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eSoil and glacial deposits may be limiting to root growth or water availability (Table\u0026nbsp;\u003cspan refid=\"Tab3\" class=\"InternalRef\"\u003e3\u003c/span\u003e). Soil attributes included tile fractional cover of 1) hydric soils with permanent or temporary oversaturation, 2) overdrained soils, 3) eroded soils, 4) densic horizon close to the surface, 5) stony soils, and 6) restrictive soils. Metrics relating to glacial deposit attributes included tile fractional cover of 1) till, 2) outwash, and 3) bedrock.\u003c/p\u003e \u003cp\u003eTopography can influence climate which may affect tree survival during stressful events. Hilltops and steep south-facing slopes may be drier than valleys or north-facing slopes. To characterize relevant slope conditions, we calculated slope and aspect based the 1 m DEM. Slope and aspect corresponded to the steepest gradient within a 3x3 window centered on each pixel (Esri, 2022). We extracted slopes\u0026thinsp;\u0026gt;\u0026thinsp;20\u0026deg; as well as slopes that were both steep (\u0026gt;\u0026thinsp;20\u0026deg;) and south-facing with aspects between 225\u0026deg; and 315\u0026deg;. Metrics were calculated as the fraction of the tiles covered by steep slopes and steep south-facing slopes (Table\u0026nbsp;\u003cspan refid=\"Tab3\" class=\"InternalRef\"\u003e3\u003c/span\u003e). To identify hilltops and valley bottoms, we used a Topographic Position Index (TPI) (Reu et al., 2013) based on the DEM. We calculated TPI by subtracting the average elevation within 20 m of a given pixel location from the actual elevation at that location. Positive TPI values indicate hilltops whereas negative values indicate valley bottoms. We extracted TPI\u0026thinsp;\u0026gt;\u0026thinsp;0.2 and TPI\u0026thinsp;\u0026lt;\u0026thinsp;0.2 to represent hilltops and valley bottoms, respectively. The fractional cover of tiles covered by hilltops and valley bottoms were used as the metrics (Table\u0026nbsp;\u003cspan refid=\"Tab3\" class=\"InternalRef\"\u003e3\u003c/span\u003e).\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab1\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eList of forest-related predictors.\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"4\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eName\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eDescription\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eJustification\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eData\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eDefoliation index\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e% defoliated forest cover weighted by number of years of defoliation.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eDefoliation depletes energy reserves.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eLandsat\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eEvergreen cover\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e% cover by coniferous trees\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eConiferous trees are not a preferred food source but do not tolerate defoliation.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eEcological Communities\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCanopy cover\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e% of 1st returns intercepted by canopy\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eCanopy cover can affect competition and microclimate.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e2011 Lidar\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eRuderal forests\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e% cover by early successional forest species\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003eEarly successional forest has species adapted to disturbances. Heavily managed forest may have different tolerances for stressors.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003eEcological Communities\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePlantations\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e% cover plantation forest\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab2\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 2\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eList of proximity- and drought-related predictors.\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"4\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eName\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eDescription\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eJustification\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eData\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eDrought index\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eAverage summer drought rating during outbreak\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eDrought is a major stressor.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eU.S. Drought Monitor\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCoast proximity\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eAverage distance to the coast\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eProximity to coast affects climate conditions.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eRI Coastal Waters\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eUrban proximity\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eAverage distance to development\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eDeveloped areas may be exposed to more pollutants and invasive species.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eNOAA CCAP land cover\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eLake proximity\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eAverage distance to the lakes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eLocations near water bodies may be associated with more abundant groundwater.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eLakes Ponds and Reservoirs\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab3\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 3\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eList of soil- and topography-related predictors.\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"4\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eName\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eDescription\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eJustification\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eData\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHydric soils\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e% cover by hydric soils\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\" morerows=\"7\" rowspan=\"8\"\u003e \u003cp\u003eHydric soils may have greater water availability whereas overdrained soils may have low available moisture. Restrictive soils, shallow bedrock, tills, and stony soils can limit root depth or lateral distribution, thus, reducing access to water and nutrients. Eroded and outwash soils may have low nutrient availability.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\" morerows=\"5\" rowspan=\"6\"\u003e \u003cp\u003eNRCS Soils\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eRestrictive soils\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e% cover by restrictive soils\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eStony soils\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e% cover by stony soils\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eShallow bedrock\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e% cover by soils with shallow bedrock\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eOverdrained soils\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e% cover by overdrained soils\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eEroded soils\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e% cover by eroded soils\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eOutwash\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e% cover by glacial outwash\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003eGlacial deposits\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTill\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e% cover by glacial till\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eValley bottoms\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e% cover by valleys (TPI\u0026thinsp;\u0026lt;\u0026thinsp;0.2)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003eValleys and depressions may have richer soils with higher moisture content compared to hilltops\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\" morerows=\"3\" rowspan=\"4\"\u003e \u003cp\u003e2011 Lidar\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHilltops\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e% cover by hilltops (TPI\u0026thinsp;\u0026gt;\u0026thinsp;0.2)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSteep slopes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e% cover by slopes\u0026thinsp;\u0026gt;\u0026thinsp;20\u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(^\\circ\\)\u003c/span\u003e\u003c/span\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003eSteeper slopes can affect soil moisture, nutrient availability, and stability. South-facing tend to be warmer and drier.\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSteep south-facing slopes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e% steep slopes facing 225\u0026deg;-315\u0026deg;\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv id=\"Sec8\" class=\"Section2\"\u003e \u003ch2\u003e2.4. Training/validation data\u003c/h2\u003e \u003cp\u003eTo train and validate our tree mortality models, we used two different \u0026ldquo;ground-truth\u0026rdquo; datasets that were based on summertime 2019 aerial imagery. The first set included 1426 tiles that were randomly selected from the 100 m polygon grid. These tiles were visually evaluated using the aerial imagery and rated using a six-category system to characterize mortality of overstory trees in each tile (Table\u0026nbsp;\u003cspan refid=\"Tab4\" class=\"InternalRef\"\u003e4\u003c/span\u003e). The second training/validation dataset included 787 tiles corresponding to Rhode Island Department of Environmental Management (RIDEM) property that had particularly high rates of tree mortality. For these tiles, the locations of dead overstory trees were digitized as point features and subsequently aggregated by tile to create ratings that were consistent with the first training dataset (Table\u0026nbsp;\u003cspan refid=\"Tab4\" class=\"InternalRef\"\u003e4\u003c/span\u003e). Altogether, the training/validation dataset included 2320 sample tiles.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab4\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 4\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eClasses of tree mortality in training/validation dataset.\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"5\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eClass\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eDead trees / ha\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eNumber of Tiles (Dataset 1)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eNumber of Tiles (Dataset 2)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003eTotal Number of Tiles\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e258\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e20\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e278\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1\u0026ndash;2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e299\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e50\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e349\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e3\u0026ndash;5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e239\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e101\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e340\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e6\u0026ndash;10\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e222\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e176\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e398\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e4\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e11\u0026ndash;20\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e152\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e272\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e424\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e\u0026gt;\u0026thinsp;20\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e8\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e523\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e531\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec9\" class=\"Section2\"\u003e \u003ch2\u003e2.5. Modeling Tree Mortality using Random Forest\u003c/h2\u003e \u003cp\u003eWe used \u003cem\u003eRandom Forest\u003c/em\u003e (Scikit-Learn, 2022) to model tree mortality based on the 21 environmental and defoliation metrics (Tables\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e\u0026ndash;\u003cspan refid=\"Tab3\" class=\"InternalRef\"\u003e3\u003c/span\u003e). We created two different models \u0026ndash; one predicted two classes of mortality (low/high) and the other predicted three classes of mortality (low/med/high). The two-class model predicted low and high mortality classes with \u0026lt;\u0026thinsp;5 and \u0026ge;\u0026thinsp;5 dead overstory trees per hectare, respectively. The three-class model predicted low, medium, and high mortality classes with \u0026le;\u0026thinsp;2, 3\u0026ndash;11, and \u0026gt;\u0026thinsp;11 dead overstory trees per hectare. We selected these thresholds to ensure there were relatively similar numbers of tiles representing each class. We used 5-fold cross-validation with a split of 80% and 20% for training and validation data, respectively, for each fold.\u003c/p\u003e \u003cp\u003eWe tuned our model by adjusting several hyperparameters through trial and error to find optimal parameters. Hyperparameters are the settings that are not learned by the model during training but set beforehand, and they can significantly impact the model's performance. We adjusted the following hyperparameters: number of trees, minimum number of samples required to be at a leaf node, and the criterion for splitting a node. We found the models performed best with 100 trees, a minimum of 3 features per node, and using the Gini criterion for splitting the nodes.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec10\" class=\"Section2\"\u003e \u003ch2\u003e2.5. Accuracy Assessment\u003c/h2\u003e \u003cp\u003eWe assessed the accuracy of our defoliation mapping for 2015, 2016 and 2017 by qualitatively assessing Landsat, aerial, and Sentinel imagery, respectively. These image datasets represented the best available data for each year. For both defoliation mapping and the mortality models, we quantitatively assessed accuracy using F1 score, precision, and recall. These metrics are based on true positives (TP), false positives (FP), and false negatives (FN) and were calculated as follows:\u003cdiv id=\"Equ3\" class=\"Equation\"\u003e\u003cdiv format=\"TEX\" class=\"mathdisplay\" id=\"FileID_Equ3\" name=\"EquationSource\"\u003e\n$$Precision=TP/(TP+FP)$$\u003c/div\u003e\u003cdiv class=\"EquationNumber\"\u003e3\u003c/div\u003e\u003c/div\u003e\u003cdiv id=\"Equ4\" class=\"Equation\"\u003e\u003cdiv format=\"TEX\" class=\"mathdisplay\" id=\"FileID_Equ4\" name=\"EquationSource\"\u003e\n$$Recall=TP/(TP+FN)$$\u003c/div\u003e\u003cdiv class=\"EquationNumber\"\u003e4\u003c/div\u003e\u003c/div\u003e\u003cdiv id=\"Equ5\" class=\"Equation\"\u003e\u003cdiv format=\"TEX\" class=\"mathdisplay\" id=\"FileID_Equ5\" name=\"EquationSource\"\u003e\n$$F1 score=2*\\frac{Precision*Recall}{Precision+Recall}$$\u003c/div\u003e\u003cdiv class=\"EquationNumber\"\u003e5\u003c/div\u003e\u003c/div\u003e\u003c/p\u003e \u003cp\u003eQuantitative assessment of the defoliation mapping was based on the 200 semi-random sample locations. Landsat, aerial, and Sentinel-2 imagery was used for visually classifying the sample points for 2015, 2016, and 2017, respectively. Accuracy assessment for the mortality modeling was based on the validation subsets (i.e. 20%) of the 2213 \u0026ldquo;ground-truth\u0026rdquo; tiles. The mortality models were further evaluated using Shapley Additive explanations (SHAP) plots and feature importance scores. SHAP plots indicate how the model uses a particular predictor and feature importance scores indicate the relative contribution of each predictor in the model classification.\u003c/p\u003e \u003c/div\u003e"},{"header":"3. Results","content":"\u003cdiv id=\"Sec12\" class=\"Section2\"\u003e \u003ch2\u003e3.1. Defoliation\u003c/h2\u003e \u003cp\u003eDuring the 3-year Spongy moth outbreak (2015\u0026ndash;2017), approximately 23% of Rhode Island's forest (312 km\u003csup\u003e2\u003c/sup\u003e) experienced significant defoliation for one or more years, based on the 30 m resolution defoliation model. Repeated defoliation for a given area was uncommon. Of the forest land that experienced defoliation during the outbreak, 89.2% of forests (278 km\u003csup\u003e2\u003c/sup\u003e) were defoliated only once. Only 10.6% (33 km\u003csup\u003e2\u003c/sup\u003e) and 0.2% (\u0026lt;\u0026thinsp;1 km\u003csup\u003e2\u003c/sup\u003e) of affected forestland experienced 2 or 3 defoliations, respectively (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e). For the approximately 136 K one-hectare study tiles with \u0026gt;\u0026thinsp;75% forest cover, 33% (44,807 tiles) and 18% (24,050 tiles) of the tiles had defoliation over more than 25% and 75% of the tile area, respectively. Defoliation was distributed predominantly in the western part of Rhode Island, with repeated defoliation occurring sporadically, mostly in the southwestern and northwestern parts of the state. Detectable defoliation rarely occurred in coastal forests (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eThe defoliation model had good agreement with our visual interpretation of the aerial and satellite imagery for our 200 sample points. For accuracy assessment, we considered 1) complete defoliation only and 2) complete and partial defoliation. We had high confidence in our ability to visually detect complete defoliation in both satellite and aerial imagery, but lower confidence in our ability to detect partial defoliation in the relatively coarse Landsat and Sentinel-2 data. For complete defoliation, the model performed consistently throughout the outbreak period with F1 scores ranging from 0.84\u0026ndash;0.86. When partial defoliation was also considered, the model accuracy was more variable with F1-scores ranging from 0.76\u0026ndash;0.87; the best accuracy was associated with the 2016 data for which we were able to use aerial imagery to create more reliable validation data.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab5\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 5\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eAccuracy assessment of defoliation model based on 200 sample locations.\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"5\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eYear\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eValidation Imagery\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eDefoliation Type\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eF1-score\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003eSample Size\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003e2015\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003eLandsat\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003ePartial\u0026thinsp;+\u0026thinsp;complete\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.79\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e113\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eComplete\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.84\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e110\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003e2016\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003eAerial\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003ePartial\u0026thinsp;+\u0026thinsp;complete\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.87\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e167\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eComplete\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.86\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e136\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003e2017\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003eSentinel\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003ePartial\u0026thinsp;+\u0026thinsp;complete\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.76\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e188\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eComplete\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.86\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e154\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec13\" class=\"Section2\"\u003e \u003ch2\u003e3.2. Tree mortality\u003c/h2\u003e \u003cp\u003eMortality was modeled for 1358 km\u003csup\u003e2\u003c/sup\u003e of forest land which corresponded to tiles with \u0026gt;\u0026thinsp;75% forest cover. Based on the models, tree mortality was highest in the western part of the state (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e). There was low tree mortality in the coastal areas and the eastern part of the state. For the 2-class model, 65% (883 km\u003csup\u003e2\u003c/sup\u003e) and 35% (475 km\u003csup\u003e2\u003c/sup\u003e) of forestland was classified as low and high mortality, respectively. For the 3-class model, 56% (760 km\u003csup\u003e2\u003c/sup\u003e), 23% (312 km\u003csup\u003e2\u003c/sup\u003e), and 21% (286 km\u003csup\u003e2\u003c/sup\u003e) of forestland was classified as low, medium, and high mortality, respectively.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eOur best 2-class and 3-class models of tree mortality had overall accuracies of 82% and 65%, respectively (Table\u0026nbsp;\u003cspan refid=\"Tab6\" class=\"InternalRef\"\u003e6\u003c/span\u003e, \u003cspan refid=\"Tab7\" class=\"InternalRef\"\u003e7\u003c/span\u003e). When only defoliation was included as a predictive variable, accuracies were 72% and 56% for the 2-class and 3-class models, respectively. When the top 3 predictive variables were included, accuracies improved to 79% and 63% for the 2-class and 3-class models, respectively. For both 2- and 3-class models, precision tended to improve slightly with the inclusion of more predictive variables. However, the 21-variable models were not consistently better than the 7-variable model, and both models were only slightly better than the 3-variable model. For the 2-class model, recall tended to improve with the number of variables. However, for the 3-class models, there was no consistent pattern between recall and number of variables.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab6\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 6\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eAccuracy assessment for \u003cb\u003e2-class models\u003c/b\u003e (low and high mortality) with varying numbers of variables. Metrics include overall accuracy (OA), cross-validation accuracy (XV), kappa (kp), F1 score (F1), precision (P), and recall (R).\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"10\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c7\" colnum=\"7\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c8\" colnum=\"8\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c9\" colnum=\"9\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c10\" colnum=\"10\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003eNumber of variables\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003eOA\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003eXV\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003ekp\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colspan=\"3\" nameend=\"c7\" namest=\"c5\"\u003e \u003cp\u003eLow\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colspan=\"3\" nameend=\"c10\" namest=\"c8\"\u003e \u003cp\u003eHigh\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003eF1\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c6\"\u003e \u003cp\u003eP\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c7\"\u003e \u003cp\u003eR\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c8\"\u003e \u003cp\u003eF1\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c9\"\u003e \u003cp\u003eP\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c10\"\u003e \u003cp\u003eR\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e21\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.81\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.78\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.61\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.77\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.82\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.73\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e0.84\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e0.81\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c10\"\u003e \u003cp\u003e0.87\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e7\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.82\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.77\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.63\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.77\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.78\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.76\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e0.86\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e0.85\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c10\"\u003e \u003cp\u003e0.86\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.79\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.77\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.56\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.73\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.77\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.70\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e0.83\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e0.80\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c10\"\u003e \u003cp\u003e0.85\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.72\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.74\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.42\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.67\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.66\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.67\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e0.75\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e0.76\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c10\"\u003e \u003cp\u003e0.75\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab7\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 7\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eAccuracy assessment for \u003cb\u003e3-class models\u003c/b\u003e (low, medium, and high mortality) with varying numbers of variables. Metrics include overall accuracy (OA), cross-validation accuracy (XV), kappa (kp), F1 score (F1), precision (P), and recall (R).\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"13\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c7\" colnum=\"7\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c8\" colnum=\"8\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c9\" colnum=\"9\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c10\" colnum=\"10\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c11\" colnum=\"11\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c12\" colnum=\"12\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c13\" colnum=\"13\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003eNumber of variables\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003eOA\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003eXV\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003ekp\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colspan=\"3\" nameend=\"c7\" namest=\"c5\"\u003e \u003cp\u003eLow\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colspan=\"3\" nameend=\"c10\" namest=\"c8\"\u003e \u003cp\u003eMedium\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colspan=\"3\" nameend=\"c13\" namest=\"c11\"\u003e \u003cp\u003eHigh\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003eF1\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c6\"\u003e \u003cp\u003eP\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c7\"\u003e \u003cp\u003eR\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c8\"\u003e \u003cp\u003eF1\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c9\"\u003e \u003cp\u003eP\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c10\"\u003e \u003cp\u003eR\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c11\"\u003e \u003cp\u003eF1\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c12\"\u003e \u003cp\u003eP\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c13\"\u003e \u003cp\u003eR\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e21\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.64\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.60\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.46\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.67\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.63\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.72\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e0.47\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e0.54\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c10\"\u003e \u003cp\u003e0.41\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c11\"\u003e \u003cp\u003e0.75\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c12\"\u003e \u003cp\u003e0.71\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c13\"\u003e \u003cp\u003e0.80\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e7\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.65\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.61\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.47\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.65\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.67\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.64\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e0.50\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e0.45\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c10\"\u003e \u003cp\u003e0.51\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c11\"\u003e \u003cp\u003e0.77\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c12\"\u003e \u003cp\u003e0.77\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c13\"\u003e \u003cp\u003e0.70\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.63\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.60\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.43\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.65\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.66\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.64\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e0.48\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e0.45\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c10\"\u003e \u003cp\u003e0.51\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c11\"\u003e \u003cp\u003e0.74\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c12\"\u003e \u003cp\u003e0.77\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c13\"\u003e \u003cp\u003e0.48\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.56\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.55\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.33\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.60\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.51\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.73\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e0.30\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e0.23\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c10\"\u003e \u003cp\u003e0.40\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c11\"\u003e \u003cp\u003e0.68\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c12\"\u003e \u003cp\u003e0.67\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c13\"\u003e \u003cp\u003e0.70\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003eThe importance of the predictive variables was consistent among all the tree mortality models (Tables\u0026nbsp;\u003cspan refid=\"Tab8\" class=\"InternalRef\"\u003e8\u003c/span\u003e, \u003cspan refid=\"Tab9\" class=\"InternalRef\"\u003e9\u003c/span\u003e, \u003cspan refid=\"Tab10\" class=\"InternalRef\"\u003e10\u003c/span\u003e). The \u003cem\u003edefoliation index\u003c/em\u003e was the most important predictor followed by \u003cem\u003ecoast proximity\u003c/em\u003e and \u003cem\u003ecanopy cover\u003c/em\u003e. For the models that included either all 21 variables or only the top 7 variables, the \u003cem\u003edefoliation index\u003c/em\u003e was substantially more important than \u003cem\u003ecoast proximity\u003c/em\u003e, and \u003cem\u003ecoast proximity\u003c/em\u003e was substantially more important than \u003cem\u003ecanopy cover\u003c/em\u003e. In models with 7 or more variables, \u003cem\u003ecanopy cover\u003c/em\u003e was only slightly more important than \u003cem\u003eevergreen cover\u003c/em\u003e. \u003cem\u003eEvergreen cover, drought index\u003c/em\u003e, \u003cem\u003eurban proximity\u003c/em\u003e, and \u003cem\u003elake proximity\u003c/em\u003e all had similar but relatively minor importance. Variables relating to soil and topography tended to be the least important predictors.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab8\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 8\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eRelative importance of predictors used for the \u003cb\u003e21-variable models\u003c/b\u003e. Higher scores indicate greater importance and all scores sum to 1.\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"3\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePredictors\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003e2-class Model\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003e3-class Model\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eDefoliation index\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.25\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.21\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCoast proximity\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.16\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.13\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCanopy cover\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.08\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.09\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eEvergreen cover\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.06\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.06\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eDrought index\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.06\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.06\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eUrban proximity\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.06\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.06\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eLake proximity\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.05\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.05\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eValley bottoms\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.04\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.04\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHilltops\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.04\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.04\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSteep slopes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.03\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.04\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHydric soils\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.02\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.03\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eStony soils\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.02\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.03\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eRestrictive soils\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.02\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.02\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eShallow bedrock\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.02\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.02\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSteep south-facing slopes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.01\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.02\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePlantations\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.01\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.01\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTill\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.01\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.01\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eOutwash\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.01\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.01\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eOver drained soils\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.01\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.01\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eEroded soils\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.01\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.01\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eRuderal forest\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.01\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.01\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab9\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 9\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eRelative importance of predictors used for the \u003cb\u003e7-variable models\u003c/b\u003e. Higher scores indicate greater importance, and all scores sum to 1.\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"3\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePredictors\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003e2-class Model\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003e3-class Model\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eDefoliation index\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.30\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.27\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCoast proximity\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.23\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.20\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCanopy cover\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.11\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.13\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eDrought index\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.09\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eEvergreen cover\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.09\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.09\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eUrban proximity\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.09\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.11\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eLake proximity\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.08\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.1\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab10\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 10\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eRelative importance of predictors used for the \u003cb\u003e3-variable models\u003c/b\u003e. Higher scores indicate greater importance, and all scores sum to 1.\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"3\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePredictors\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003e2-class Model\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003e3-class Model\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eDefoliation index\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.38\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.37\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCoast proximity\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.36\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.36\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCanopy cover\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.27\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.28\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003eThe SHAP analysis shows how individual variables influence the model results for each predicted class. For both the 2- and 3-class models, tiles that were classified as \u0026ldquo;low mortality\u0026rdquo; tended to have lower \u003cem\u003edefoliation index\u003c/em\u003e, closer \u003cem\u003ecoast proximity\u003c/em\u003e, higher \u003cem\u003ecanopy cover\u003c/em\u003e, lower \u003cem\u003edrought index\u003c/em\u003e, closer \u003cem\u003eurban proximity\u003c/em\u003e, and higher \u003cem\u003eevergreen cover\u003c/em\u003e. Conversely, tiles that were classified as \u0026ldquo;high mortality\u0026rdquo; tended to have higher \u003cem\u003edefoliation index\u003c/em\u003e, further \u003cem\u003ecoast proximity\u003c/em\u003e, lower \u003cem\u003ecanopy cover\u003c/em\u003e, higher \u003cem\u003edrought index\u003c/em\u003e, further \u003cem\u003eurban proximity\u003c/em\u003e, and less \u003cem\u003eevergreen cover\u003c/em\u003e.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e"},{"header":"4. Discussion","content":"\u003cp\u003eThis study used satellite-based defoliation mapping combined with geospatial environmental predictors to model tree mortality resulting from a Spongy moth outbreak in a mixed temperate deciduous forest. The models identified predictors that were associated with higher rates of mortality (e.g. \u003cem\u003edefoliation index, coast proximity, canopy cover\u003c/em\u003e). The inclusion of environmental predictors along with the \u003cem\u003edefoliation index\u003c/em\u003e resulted in significant improvements to model performance. This finding is consistent with Campbell et al. (2020) found that GIS-based data explained an additional 17% of mortality beyond defoliation data alone for their semi-arid woodland study area.\u003c/p\u003e \u003cp\u003eThe performance of our models was on par with previous studies which focused on coniferous forests with relatively lower species diversity. Meddens et al. (2014), Goodwin et al. (2008), and Long et al. (2016) achieved accuracies of 70\u0026ndash;80% for their coniferous forest study areas. Bergm\u0026uuml;ller et al. (2022) achieved an accuracy of 70% in Canadian mixed forests using very high-resolution imagery collected by unmanned aerial systems (UAS). Using GIS-based topographic data and field-based soil samples, Toledo et al. (2011) were able to explain 20% of tree mortality. Campbell et al. (2020) was one of the few studies that combined a defoliation metric with GIS-based environmental factors to model tree mortality. For their open-canopy semi-arid woodland, their model achieved an r\u003csup\u003e2\u003c/sup\u003e\u0026thinsp;=\u0026thinsp;0.71.\u003c/p\u003e \u003cp\u003eThe SHAP analysis was generally consistent with our expectations of how various predictors should affect tree mortality. Higher tree mortality was associated with a higher \u003cem\u003edefoliation index\u003c/em\u003e, further distance from coast, lower canopy cover, less evergreen cover, and closer proximity to urban cover. Since defoliation is the primary stressor for trees during a Spongy moth outbreak, the \u003cem\u003edefoliation index\u003c/em\u003e was expected to be directly related to increased mortality. \u003cem\u003eEvergreen cover\u003c/em\u003e was expected to be inversely related to mortality since coniferous species are not the preferred food source of Spongy moths. The higher mortality further from the coast may be due to higher temperatures and lower humidity found in inland areas (NCDC, 1971). Coastal proximity was much more important than drought index which suggests that other protective attributes (e.g. such as differing forest compositions) may be associated with coastal areas. The much higher resolution of the coastal proximity metric may also help explain its greater importance than the drought index. The higher mortality associated with lower canopy cover may be due to the increased solar heating of the ground and evaporation of soil moisture in more open forests. In our study area, forests with lower canopy cover could also signify previous disturbances that contributed to overall poor forest health. Our finding regarding canopy cover is consistent with Campbell et al. (2020) who studied semi-arid pi\u0026ntilde;on-juniper woodlands. The protective effect of closer proximity to urban areas was somewhat unexpected. However, this relatively minor effect may be due to the greater care and management (e.g. prompt removal of dead trees) provided to trees in more urban environments.\u003c/p\u003e \u003cp\u003eWe unexpectedly found that soil-based predictors had little importance in our models for predicting tree mortality. Adverse soil conditions constrain maximum tree heights, slow growth rates, and stress trees through limited water or nutrient availability. Campbell et al. (2020) found that surface organic matter had moderate importance for predicting tree mortality but found that 29 other soil variables had very little value. The lack of importance for soils data in our study may reflect a limitation of GIS soils datasets. The minimum mapping unit of our dataset was around 1 ha which could omit much of the soil variation that would be relevant to mortality of individual trees. In addition, unfavorable individual soil characteristics were relatively uncommon in our study area and the majority of the tiles in our training/validation dataset had zeroes or very low values for soil-based predictors. The lack of variation may have made it less likely for \u003cem\u003eRandom Forest\u003c/em\u003e to find useful partitions of these predictors associated with varying levels of tree mortality. Combining the soil characteristics into a single metric may yield a more useful predictor for Random Forest.\u003c/p\u003e \u003cp\u003eWe found that topographic characteristics also had little importance in our models of tree mortality. Steeper south-facing slopes tend to receive more solar heating than other slope orientations which results in warmer and drier conditions that are likely to stress trees. However, the topography in our study area was relatively moderate with little area covered by steep south-facing slopes. The relatively infrequent occurrence of steep slopes in the study area may have made the predictor unlikely to be used effectively by \u003cem\u003eRandom Forest\u003c/em\u003e. However, slope and orientation may be more important factors in areas with rugged topography.\u003c/p\u003e \u003cp\u003eOur models were only slightly improved when we included more than the 3 top predictors. Models with the 7 top predictors performed very similarly to models with the full set of predictors. The ability to use fewer predictors without sacrificing model performance is advantageous because it simplifies model development and improves efficiency.\u003c/p\u003e \u003cp\u003eThe fraction of a tree crown that is defoliated is likely to be an important factor in tree mortality. However, our defoliation metric was based primarily on complete defoliation (i.e. ~100% crown defoliation). Pasquarella et al. (2018) did map defoliation, from a Spongy moth outbreak, with differing levels of severity using a Tasseled Cap transform approach. However, we did not attempt to map differing severity levels because we found it difficult to assess partial defoliation visually based on the 10\u0026ndash;30 m resolutions of the satellite imagery that we used for accuracy assessments. We also found significant annual variation in our baseline (i.e. pre-outbreak) NDVI values; thus, we chose a conservative threshold for defoliation to minimize commission error.\u003c/p\u003e \u003cp\u003eThis study was applied to a Spongy moth pest outbreak over a limited geographic area. Thus, the relevant predictors of mortality may be somewhat different for other forest pests or disease outbreaks in different areas. However, we did confirm the benefit of incorporating GIS-based predictors along with satellite-based mapping of defoliation in forest mortality models in temperate deciduous forests. The \u003cem\u003eRandom Forest\u003c/em\u003e tool was effective for using certain types of predictors in ways that are consistent with expectations. However, it seemed ineffective for using predictors that are relatively uncommon but still likely to be relevant (e.g. soils). Limitations in the spatial resolution of GIS data may also preclude the inclusion of important predictors in mortality models. Future work should explore whether uncommon, but likely relevant, features can be used more effectively in mortality models.\u003c/p\u003e"},{"header":"5. Conclusion","content":"\u003cp\u003eOur study used satellite-based defoliation mapping and geospatial environmental data with \u003cem\u003eRandom Forest\u003c/em\u003e to model tree mortality from a Spongy moth outbreak in Rhode Island\u0026rsquo;s temperate deciduous forest. The best models achieved accuracies of 82% and 65% when predicting 2 classes (low/high) or 3 classes (low/med/high) of mortality, respectively. The inclusion of geospatial data improved model predictions by 7\u0026ndash;10% compared to models based only on defoliation. The most important predictors in the models were the \u003cem\u003edefoliation index\u003c/em\u003e, \u003cem\u003ecoastal proximity\u003c/em\u003e, and \u003cem\u003ecanopy cover\u003c/em\u003e. Models improved only slightly with the inclusion of more than 3 top predictors. Soil characteristics had very little contribution to the models which may be due to the coarse resolution (i.e. minimum mapping unit of 1 ha) and the tendency for individual adverse soil characteristics to be relatively uncommon. Topographic factors also had minimal influence on models which may be due to the relatively moderate topography of the study area. Although relevant predictors of tree mortality may vary somewhat for different regions and pest species, this study showed the benefit of \u003cem\u003eRandom Forest\u003c/em\u003e modeling that combines satellite-based monitoring with geospatial environmental data.\u003c/p\u003e"},{"header":"Declarations","content":"\u003ch2\u003eAuthor Contribution\u003c/h2\u003e\u003cp\u003eJason Parent and Liubov Dumarevskaya jointly conceived and designed the study. Liubov Dumarevskaya conducted data collection and analysis under Jason Parent's supervision. Liubov Dumarevskaya drafted the manuscript, and Jason Parent provided critical revisions for important intellectual content. Both authors approved the final version of the manuscript for submission.\u003c/p\u003e\u003ch2\u003eAcknowledgement\u003c/h2\u003e\u003cp\u003eThe authors thank Rockwell Richards and Molly Ahern for their contributions in creating the training/validation datasets used in this study.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eAnderegg, W. R. L., Hicke, J. A., Fisher, R. A., Allen, C. D., Aukema, J., Bentz, B., Hood, S., Lichstein, J. W., Macalady, A. K., McDowell, N., Pan, Y., Raffa, K., Sala, A., Shaw, J. D., Stephenson, N. L., Tague, C., Zeppel, M. 2015. Tree mortality from drought, insects, and their interactions in a changing climate. New Phytologist, 208(3), 674\u0026ndash;683. DOI: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1111/nph.13477\u003c/span\u003e\u003cspan address=\"10.1111/nph.13477\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBaker, W. L. 1941. Effect of Gypsy Moth Defoliation on Certain Forest Trees. Journal of Forestry, Volume 39, Issue 12, Pages 1017\u0026ndash;1022. DOI: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1093/jof/39.12.1017\u003c/span\u003e\u003cspan address=\"10.1093/jof/39.12.1017\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBarbosa, P., Capinera, J.L., 1978. Population quality, dispersal and numerical change in the gypsy moth, Lymantria dispar (L.). Oecologia, 36, pp.203\u0026ndash;209. DOI: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1007/BF00349809\u003c/span\u003e\u003cspan address=\"10.1007/BF00349809\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBergm\u0026uuml;ller, K. O., Vanderwel, M. C. 2022. Predicting Tree Mortality Using Spectral Indices Derived from Multispectral UAV Imagery. Remote Sensing, 14(9), 2195. DOI: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.3390/rs14092195\u003c/span\u003e\u003cspan address=\"10.3390/rs14092195\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBigsby, K.M., Ambrose, M.J., Tobin, P.C. Sills, E.O., 2014. The cost of gypsy moth sex in the city. Urban Forestry \u0026amp; Urban Greening, 13(3), pp.459\u0026ndash;468. DOI: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.ufug.2014.05.003\u003c/span\u003e\u003cspan address=\"10.1016/j.ufug.2014.05.003\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBottero, A., D'Amato, A.W., Palik, B.J., Bradford, J.B., Fraver, S., Battaglia, M.A., Asherin, L.A., 2017. Density-dependent vulnerability of forest ecosystems to drought. Journal of Applied Ecology, 54(6), pp.1605\u0026ndash;1614. DOI: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1111/1365-2664.12847\u003c/span\u003e\u003cspan address=\"10.1111/1365-2664.12847\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCampbell, R. W. 1979. Gypsy Moth Influence on Forest. Agriculture Information Bulletin No. 423. United States Department of Agriculture, Forest Service, Pacific Northwest Forest and Range Experiment Station.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCampbell, M. J., Dennison, P. E., Tune, J. W., Kannenberg, S. A., Kerr, K. L., Codding, B. F., Anderegg, W. R. L. 2020. A multi-sensor, multi-scale approach to mapping tree mortality in woodland ecosystems. Remote Sensing of Environment, Volume 245, 111853. DOI: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.rse.2020.111853\u003c/span\u003e\u003cspan address=\"10.1016/j.rse.2020.111853\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eChoat, B., Brodribb, T.J., Brodersen, C.R., Duursma, R.A., L\u0026oacute;pez, R. and Medlyn, B.E., 2018. Triggers of tree mortality under drought. Nature, 558(7711), pp.531\u0026ndash;539. DOI: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1038/s41586-018-0240-x\u003c/span\u003e\u003cspan address=\"10.1038/s41586-018-0240-x\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNational Oceanographic and Atmospheric Administration (NOAA). 2020. Coastal Change Analysis Program (C-CAP). High-Resolution Land Cover dataset for Rhode Island. Retrieved from [\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://chs.coast.noaa.gov/htdata/raster1/landcover/bulkdownload/hires/ri/]\u003c/span\u003e\u003cspan address=\"https://chs.coast.noaa.gov/htdata/raster1/landcover/bulkdownload/hires/ri/]\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e on July 2022.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDas, A., Battles, J., Stephenson, N. L., van Mantgem, P. J. 2011. The contribution of competition to tree mortality in old-growth coniferous forests. Forest Ecology and Management, 261(7), 1203\u0026ndash;1213.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDavidson, C. B., Gottschalk, K. W., \u0026amp; Johnson, J. E. 1999. Tree mortality following defoliation by the European gypsy moth (Lymantria dispar L.) in the United States: a review. Forest science, 45(1), 74\u0026ndash;84.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDorman, M., Perevolotsky, A., Sarris, D., Svoray, T. 2015. The effect of rainfall and competition intensity on forest response to drought: lessons learned from a dry extreme. Oecologia, 177, 1025\u0026ndash;1038. DOI: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1007/s00442-015-3229-2\u003c/span\u003e\u003cspan address=\"10.1007/s00442-015-3229-2\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDymond, J.R., 2010. Soil erosion in New Zealand is a net sink of CO\u003csup\u003e2\u003c/sup\u003e. Earth Surface Processes and Landforms, 35(15), pp.1763\u0026ndash;1772. DOI: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1002/esp.2014\u003c/span\u003e\u003cspan address=\"10.1002/esp.2014\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDunn, C.P., Stearns, F., 1987. A comparison of vegetation and soils in floodplain and basin forested wetlands of southeastern Wisconsin. American Midland Naturalist, pp.375\u0026ndash;384.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGoodwin, N. R., Coops, N. C., Wulder, M. A., Gillanders, S., Schroeder, T. A., Nelson, T. 2008. Estimation of insect infestation dynamics using a temporal sequence of Landsat data. Remote Sensing of Environment, Volume 112, Issue 9, Pages 3680\u0026ndash;3689. DOI: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.rse.2008.05.005\u003c/span\u003e\u003cspan address=\"10.1016/j.rse.2008.05.005\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGottschalk, K. W., Colbert, J. J., \u0026amp; Feicht, D. L. 2007. Tree mortality risk of oak due to gypsy moth. European Journal of Forest Pathology. DOI: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1111/j.1439-0329.1998.tb01173.x\u003c/span\u003e\u003cspan address=\"10.1111/j.1439-0329.1998.tb01173.x\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGuar\u0026iacute;n, A., \u0026amp; Taylor, A. H. 2005. Drought triggered tree mortality in mixed conifer forests in Yosemite National Park, California, USA. Forest Ecology and Management, 218(1\u0026ndash;3), 229\u0026ndash;244. DOI: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.foreco.2005.07.014\u003c/span\u003e\u003cspan address=\"10.1016/j.foreco.2005.07.014\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGuggenmoos, S. 2003. Effects of tree mortality on power line security. Journal of Arboriculture, 29(4), 181\u0026ndash;196. DOI: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.48044/jauf.2003.022\u003c/span\u003e\u003cspan address=\"10.48044/jauf.2003.022\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGunst, K., Weisberg, P.G., Yang, J., Fan. Y. Do denser forests have greater risk of tree mortality: A remote sensing analysis of density-dependent forest mortality, Forest Ecology and Management, Volume 359, 2016, Pages 19\u0026ndash;32, ISSN 0378\u0026ndash;1127, DOI: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.foreco.2015.09.032\u003c/span\u003e\u003cspan address=\"10.1016/j.foreco.2015.09.032\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eEnser, R., Gregg, D., Sparks, C., August, P., Jordan, P., Coit, J., Raithel, C., Tefft, B., Payton, B., Brown, C., LaBash, C., Comings, S., \u0026amp; Ruddock, K. 2011. Rhode Island Ecological Communities Classification. Technical Report. Rhode Island Natural History Survey, Kingston, RI. 33 pp.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eEr\u0026auml;j\u0026auml;\u0026auml;, S., Halme, P., Kotiaho, J.S., Markkanen, A., Toivanen, T., 2010. The volume and composition of dead wood on traditional and forest fuel harvested clear-cuts. Silva Fennica, 44(2). DOI:\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.14214/sf.150\u003c/span\u003e\u003cspan address=\"10.14214/sf.150\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFei, S., Morin, R. S., Oswalt, C. M., \u0026amp; Liebhold, A. M. 2019. Biomass losses resulting from insect and disease invasions in US forests. Proceedings of the National Academy of Sciences, 116(35), 17371\u0026ndash;17376. Published on August 12, 2019. DOI: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1073/pnas.1820601116\u003c/span\u003e\u003cspan address=\"10.1073/pnas.1820601116\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFranklin, J.F., Shugart, H.H. and Harmon, M.E., 1987. Tree death as an ecological process. BioScience, 37(8), pp.550\u0026ndash;556. DOI:\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.2307/1310665\u003c/span\u003e\u003cspan address=\"10.2307/1310665\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHorsley, S. B., Long, R. P., Bailey, S. W., Hallett, R. A., \u0026amp; Wargo, P. M. Health of eastern North American sugar maple forests and factors affecting decline. Northern Journal of Applied Forestry 19.1. 2002: 34\u0026ndash;44. DOI: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1093/njaf/19.1.34\u003c/span\u003e\u003cspan address=\"10.1093/njaf/19.1.34\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHudgins, E.J., Liebhold, A.M. and Leung, B., 2017. Predicting the spread of all invasive forest pests in the United States. Ecology letters, 20(4), pp.426\u0026ndash;435. DOI: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1111/ele.12741\u003c/span\u003e\u003cspan address=\"10.1111/ele.12741\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eIvantsova, E.D., Pyzhev, A.I. and Zander, E.V., 2019. Economic consequences of insect pests outbreaks in boreal forests: A literature review. Journal of Sibiric Federal University, 12(4), pp.627\u0026ndash;642. DOI: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.17516/1997-1370-0417\u003c/span\u003e\u003cspan address=\"10.17516/1997-1370-0417\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKayama, M., Quoreshi, A.M., Kitaoka, S., Kitahashi, Y., Sakamoto, Y., Maruyama, Y., Kitao, M. and Koike, T., 2003. Effects of deicing salt on the vitality and health of two spruce species, \u003cem\u003ePicea abies\u003c/em\u003e Karst., and \u003cem\u003ePicea glehnii\u003c/em\u003e Masters planted along roadsides in northern Japan. Environmental pollution, 124(1), pp.127\u0026ndash;137. DOI: DOI: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/s0269-7491(02)00415-3\u003c/span\u003e\u003cspan address=\"10.1016/s0269-7491(02)00415-3\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKinahan, I.G., Grandstaff, G., Russell, A., Rigsby, C.M., Casagrande, R.A. and Preisser, E.L., 2020. A four-year, seven-state reforestation trial with eastern hemlocks (\u003cem\u003eTsuga canadensis\u003c/em\u003e) resistant to hemlock woolly adelgid (\u003cem\u003eAdelges tsugae\u003c/em\u003e). Forests, 11(3), p.312. DOI: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.3390/f11030312\u003c/span\u003e\u003cspan address=\"10.3390/f11030312\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eUnited States Geological Survey (USGS). Available at \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://earthexplorer.usgs.gov/\u003c/span\u003e\u003cspan address=\"https://earthexplorer.usgs.gov/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (last accessed on November 2022).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLong, J. A., Lawrence, R. L. 2016. Mapping Percent Tree Mortality Due to Mountain Pine Beetle Damage. Forest Science, Volume 62, Issue 4, Pages 392\u0026ndash;402. DOI: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.5849/forsci.15-046\u003c/span\u003e\u003cspan address=\"10.5849/forsci.15-046\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMa, Q., Su, Y., \u0026amp; Guo, Q. 2017. Comparison of Canopy Cover Estimations From Airborne LiDAR, Aerial Imagery, and Satellite Imagery. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 10(9), 4225\u0026ndash;4236. DOI: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1109/JSTARS.2017.2711482\u003c/span\u003e\u003cspan address=\"10.1109/JSTARS.2017.2711482\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMeddens, A. J. H., Hicke, J. A., Vierling, L. A., Hudak, A. T. 2013. Evaluating methods to detect bark beetle-caused tree mortality using single-date and multi-date Landsat imagery. Remote Sensing of Environment, Volume 132, Pages 49\u0026ndash;58. DOI: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.rse.2013.01.002\u003c/span\u003e\u003cspan address=\"10.1016/j.rse.2013.01.002\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMeddens, A. J. H., Hicke, J. A. 2014. Spatial and temporal patterns of Landsat-based detection of tree mortality caused by a mountain pine beetle outbreak in Colorado, USA. Forest Ecology and Management, Volume 322, Pages 78\u0026ndash;88. DOI: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.foreco.2014.02.037\u003c/span\u003e\u003cspan address=\"10.1016/j.foreco.2014.02.037\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMeng, R., Gao, R., Zhao, F., Huang, C., Sun, R., Lv, Z., Huang, Z. 2022. Landsat-based monitoring of southern pine beetle infestation severity and severity change in a temperate mixed forest. Remote Sensing of Environment, 269, 112847. DOI: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.rse.2021.112847\u003c/span\u003e\u003cspan address=\"10.1016/j.rse.2021.112847\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMurray, B.C., Pendleton, L., Jenkins, W.A. and Sifleet, S., 2011. Green payments for blue carbon: economic incentives for protecting threatened coastal habitats. Nicholas Institute for Environmental Policy Solutions, Duke University, Durham.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNavarro-Cerrillo, R.M., Varo-Mart\u0026iacute;nez, M.\u0026Aacute;., Acosta, C., Rodriguez, G.P., S\u0026aacute;nchez-Cuesta, R., Ruiz G\u0026oacute;mez, F.J. 2019. Integration of WorldView-2 and airborne laser scanning data to classify defoliation levels in Quercus ilex L. Dehesas affected by root rot mortality: Management implications. Forest Ecology and Management, 451, 117564. DOI: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.foreco.2019.117564\u003c/span\u003e\u003cspan address=\"10.1016/j.foreco.2019.117564\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNational Climatic Data Center (NCDC). 1971. Climates of the States, Volume 1. Distributed by the U.S. Govt. Print. Off., Washington. (Climatography of the United States). Retrieved from \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://books.google.com/books?id=pfHFwgEACAAJ\u003c/span\u003e\u003cspan address=\"https://books.google.com/books?id=pfHFwgEACAAJ\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNational Drought Mitigation Center. (2015\u0026ndash;2017). U.S. Drought Monitor (USDM): Weekly ratings of drought across the U.S for the summer period.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003e\u0026Oslash;ystein H. Opedal, W. Scott Armbruster \u0026amp; Bente J. Graae. 2015. Linking small-scale topography with microclimate, plant species diversity and intra-specific trait variation in an alpine landscape, Plant Ecology \u0026amp; Diversity, 8:3, 305\u0026ndash;315, DOI: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1080/17550874.2014.987330\u003c/span\u003e\u003cspan address=\"10.1080/17550874.2014.987330\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eParotta, J., Knowles, O., Wunderle, J.M., 1997. Floristic diversity development in a 10-year-old restoration forest on a bauxite mined site in Amazonia. Forest Ecol. Manage, 99, pp.21\u0026ndash;42. DOI: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/S0378-1127(97)00192-8\u003c/span\u003e\u003cspan address=\"10.1016/S0378-1127(97)00192-8\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePasquarella, V.J., Elkinton, J.S. \u0026amp; Bradley, B.A. Extensive gypsy moth defoliation in Southern New England characterized using Landsat satellite observations. Biol Invasions 20, 3047\u0026ndash;3053. 2018. DOI: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1007/s10530-018-1778-0\u003c/span\u003e\u003cspan address=\"10.1007/s10530-018-1778-0\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePercy, K.E., Ferretti, M., 2004. Air pollution and forest health: toward new monitoring concepts. Environmental pollution, 130(1), pp.113\u0026ndash;126.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePoulos, H.M. and Camp, A.E., 2010. Decision support for mitigating the risk of tree induced transmission line failure in utility rights-of-way. Environmental management, 45, pp.217\u0026ndash;226. DOI: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1007/s00267-009-9422-5\u003c/span\u003e\u003cspan address=\"10.1007/s00267-009-9422-5\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eQuirion, B.R., Domke, G.M., Walters, B.F., Lovett, G.M., Fargione, J.E., Greenwood, L., Serbesoff-King, K., Randall, J.M. and Fei, S., 2021. Insect and disease disturbances correlate with reduced carbon sequestration in forests of the contiguous United States. Frontiers in Forests and Global Change, 4, p.716582. DOI: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.3389/ffgc.2021.716582\u003c/span\u003e\u003cspan address=\"10.3389/ffgc.2021.716582\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRamsfield, T.D., Bentz, B.J., Faccoli, M., Jactel, H. and Brockerhoff, E.G., 2016. Forest health in a changing world: effects of globalization and climate change on forest insect and pathogen impacts. Forestry, 89(3), pp.245\u0026ndash;252. DOI: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1093/forestry/cpw018\u003c/span\u003e\u003cspan address=\"10.1093/forestry/cpw018\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDe Reu, J., Bourgeois, J., Bats, M., Zwertvaegher, A., Gelorini, V., De Smedt, P., Chu, W., Antrop, M., De Maeyer, P., Finke, P., Van Meirvenne, M., Verniers, J., Cromb\u0026eacute;, P. 2013. Application of the topographic position index to heterogeneous landscapes. Catena, 103, 31\u0026ndash;39. DOI: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.catena.2012.11.008\u003c/span\u003e\u003cspan address=\"10.1016/j.catena.2012.11.008\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRhode Island Geographic Information Systems (RIGIS). 2023. Rhode Island Maps and Data Geospatial Hub. Available at www.rigis.org (last accessed on November 2023).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRullan-Silva, C., Olthoff, A., Delgado de la Mata, J., \u0026amp; Pajares-Alonso, J. (2013). Remote Monitoring of Forest Insect Defoliation - A Review -. Forest Systems, 22(3), 377\u0026ndash;391. DOI: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.5424/fs/2013223-04417\u003c/span\u003e\u003cspan address=\"10.5424/fs/2013223-04417\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSeidl, R., Schelhaas, M.J., Rammer, W. and Verkerk, P.J., 2014. Increasing forest disturbances in Europe and their impact on carbon storage. Nature climate change, 4(9), pp.806\u0026ndash;810. DOI: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1038/nclimate2318\u003c/span\u003e\u003cspan address=\"10.1038/nclimate2318\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eShearman, T. M., Varner, J. M., Hood, S. M., Cansler, C. A., Hiers, J. K. 2019. Modelling post-fire tree mortality: Can random forest improve discrimination of imbalanced data? Ecological Modelling, Volume 414, 108855. DOI: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.ecolmodel.2019.108855\u003c/span\u003e\u003cspan address=\"10.1016/j.ecolmodel.2019.108855\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eScikit-Learn: Machine Learning in Python. Random Forests in Python. Retrieved from \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://scikit-learn.org/stable/modules/generated/sklearn.ensemble.RandomForestClassifier.html\u003c/span\u003e\u003cspan address=\"https://scikit-learn.org/stable/modules/generated/sklearn.ensemble.RandomForestClassifier.html\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e on November 2022.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eStravinskienė, V., Bartkevičius, E., Abraitienė, J., Dautartė, A. 2018. Assessment of Pinus sylvestris L. tree health in urban forests at highway sides in Lithuania. Global Ecology and Conservation, 16, e00517. DOI: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.gecco.2018.e00517\u003c/span\u003e\u003cspan address=\"10.1016/j.gecco.2018.e00517\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ede Toledo, J. J., Magnusson, W. E., Castilho, C. V., \u0026amp; Nascimento, H. E. M. 2012. Tree mode of death in Central Amazonia: Effects of soil and topography on tree mortality associated with storm disturbances. Forest Ecology and Management, Volume 263, Pages 253\u0026ndash;261. DOI: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.foreco.2011.09.017\u003c/span\u003e\u003cspan address=\"10.1016/j.foreco.2011.09.017\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTrumbore, S., Brando, P., Hartmann, H., 2015. Forest health and global change. Science, 349(6250), pp.814\u0026ndash;818. DOI: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1126/science.aac6759\u003c/span\u003e\u003cspan address=\"10.1126/science.aac6759\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eVerbesselt, J., Robinson, A., Stone, C. and Culvenor, D., 2009. Forecasting tree mortality using change metrics derived from MODIS satellite data. Forest Ecology and Management, 258(7), pp.1166\u0026ndash;1173. DOI: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.foreco.2009.06.011\u003c/span\u003e\u003cspan address=\"10.1016/j.foreco.2009.06.011\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhan, Z., Yu, L., Li, Z., Ren, L., Gao, B., Wang, L., Luo, Y. 2020. Combining GF-2 and Sentinel-2 Images to Detect Tree Mortality Caused by Red Turpentine Beetle during the Early Outbreak Stage in North China. Forests, 11(2), 172. DOI: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.3390/f11020172\u003c/span\u003e\u003cspan address=\"10.3390/f11020172\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhang, X., Jiao, J. J., \u0026amp; Guo, W. 2022. How Does Topography Control Topography-Driven Groundwater Flow? Geophysical Research Letters, 49(20), e2022GL101005. DOI: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1029/2022GL101005\u003c/span\u003e\u003cspan address=\"10.1029/2022GL101005\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eUnited States Census Bureau (USCB). 2020. Historical Population Density Data (1910\u0026ndash;2020). Retrieved from \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.census.gov/data/tables/time-series/dec/density-data-text.html\u003c/span\u003e\u003cspan address=\"https://www.census.gov/data/tables/time-series/dec/density-data-text.html\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e on November 2022.\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"","lastPublishedDoi":"10.21203/rs.3.rs-4378454/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-4378454/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"Invasive pests cause major ecological and economic damages to forests around the world. Satellite imagery is an important tool for monitoring defoliation at broad scales, but environmental conditions can affect whether defoliation leads to mortality. In this study, we modeled forest mortality resulting from a 2015-2017 Spongy moth outbreak in the temperate deciduous forests of Rhode Island (northeastern U.S.). We used Landsat-based defoliation mapping and geospatial environmental data with Random Forest to model mortality severity of canopy trees at a 100 m spatial resolution.\nDefoliation was mapped based on declines in Normalized Difference Vegetation Index (NDVI) in outbreak years compared to baseline NDVI from pre-outbreak years. Other predictors included geospatial data representing soil characteristics, drought condition, and forest characteristics as well as proximity to coast, development, and water. The Random Forest tool in Python Sklearn was used to model forest mortality with 2 classes (low/high) and 3 classes (low/med/high). The best models had overall accuracies of 82% and 65% for the 2-class and 3-class models, respectively. The most important predictors of forest mortality were defoliation, distance to coast, and canopy cover. Soils and topography had minimal importance in the models possibly due to limitations of the data and limited variability within our study area. Repeated defoliations were relatively rare during the outbreak. Model performance improved only slightly with the inclusion of more than 3 variables. The models classified 35% of forests as having canopy mortality \u003e 5 trees/ha and 21% of Rhode Island forests having mortality \u003e11 trees/ha. The study shows the benefit of Random Forest models that use both defoliation maps and geospatial environmental data for classifying forest mortality.","manuscriptTitle":"Modeling Spongy Moth Forest Mortality in Rhode Island Temperate Deciduous Forest","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2024-05-14 18:58:18","doi":"10.21203/rs.3.rs-4378454/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"43c6e36c-5883-4333-adb2-9d97775b7dfb","owner":[],"postedDate":"May 14th, 2024","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[],"tags":[],"updatedAt":"2024-05-20T14:00:14+00:00","versionOfRecord":[],"versionCreatedAt":"2024-05-14 18:58:18","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-4378454","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-4378454","identity":"rs-4378454","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}