Estimating canopy cover using a Spatially Balanced Sampling approach: A case study of miombo woodlands in Western Tanzania

Estimating canopy cover using a Spatially Balanced Sampling approach: A case study of miombo woodlands in Western Tanzania | 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 Estimating canopy cover using a Spatially Balanced Sampling approach: A case study of miombo woodlands in Western Tanzania Siwa E. Nkya, Deo D. Shirima, Henrik Hedenas, August B. Temu This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-3880805/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 Long-term monitoring is essential to understand the impacts of land use and climate change on miombo woodlands. This study introduces an innovative monitoring design for miombo woodlands with a two-stage sampling utilizing spatially balanced techniques to estimate the area and canopy cover of miombo woodland across the Tabora, Sikonge, Mlele, and Tanganyika districts. The first step involved the selection of 68 tracts, each comprising an average of 1025 plots, with the aid of spatially balanced sampling. Each of the 69,716 plots was classified into closed (canopy cover > 70%), open (40% ≤ canopy cover ≤ 70%), very open (10% ≤ canopy cover < 40%), and non-miombo (canopy cover < 10%) based on woodland cover derived from Sentinel 2 images, followed by the second step consisting of stratified random sampling and inventorying of 2,690 plots within 68 tracts. Using PlanetScope images, we determined the canopy cover for the 2,690 plots selected in the second step and reclassified them accordingly. Employing the Horvitz–Thompson estimator, our results showed that miombo woodlands in these districts cover 37,359 ± 4,618 km² with an average canopy cover of 55% ± 5%. Closed miombo woodland (canopy cover > 70%) was the dominating woodland type, covering 29,546 ± 4,382 km² of the study area with an average canopy cover of 84% ± 7%. The study's innovative sampling design provides reliable estimates of the area of miombo woodlands and average canopy cover, with relative standard errors consistently below 25%, offering a robust foundation for monitoring different miombo types. Two-Stage sampling Stratified Random Sampling Auxiliary variables Horvitz-Thompson estimator Remote sensing Figures Figure 1 Figure 2 Introduction Miombo refers to woodlands dominated by Brachystegia, Julbernadia and/or isoberlinia genera from the legume family - Fabaceae, subfamily Caesalpinioideae. These woodlands are extensive tropical African formations covering 2.7 million km 2 within Angola, Malawi, Mozambique, Tanzania, Zimbabwe, Zambia, and the Democratic Republic of Congo (Frost, 1996 ). In Tanzania, these woodlands cover 447263 km 2 , constituting about 93% of the total forested area (URT, 2015 ). Miombo woodlands can be categorized into dry miombo, in areas receiving annual rainfall of less than 1000 mm, and wet miombo, in areas receiving yearly rainfall equal to or above 1000 mm (Frost, 1996 ). Due to their extensiveness, miombo woodlands play a major role in poverty alleviation by providing household energy, timber, and food like mushrooms, fruits, and honey (Abdallah & Monela, 2007 ; Dewees et al., 2011 ; Lusambo, 2009 ; Syampungani et al., 2009 ). Miombo and mopane (another dry woodland in Southern Africa) harbors 8,500 plant species, of which 4,600 are endemic (Mittermeier et al., 2003 ). In recent years, their role in carbon sequestration has also been recognized (Lupala et al., 2014 ). The increasing demand for forest products and the expansion of cropland due to population increases result in woodland loss, thereby threatening the sustainable provision of these goods and services (Lusambo, 2009 ; Syampungani et al., 2009 ; URT, 2015 ). Moreover, forest loss contributes to global warming and climate change (Corbera et al., 2010 ). John et al. ( 2020 ) predicted that climate change will likely result in the loss of 5% of miombo woodland suitable habitat by 2085 due to a drier climate. However, the most significant impact might not be the net loss but rather the projected conversion of wet to dry miombo woodland. Jinga and Palagi ( 2020 ) predicted a decline in the areal extent of wet miombo by up to 41.6% by 2070, while the areal extent of dry miombo might increase by 22.7% over the same period. Thus, tree species adapted to drier conditions will be increased, replacing those miombo species thriving in wetter conditions. Since dry miombo woodlands exhibit lower tree growth rates than wet miombo (Chidumayo 2019 ), carbon sequestration may be further reduced. Moreover, even if there is no shift in tree species composition, climate change could impact the woodland canopy. Canopy structures are adapted to the maximum water deficits in the tropics (Pfeifer et al., 2018 ), and the canopy phenology in the miombo is influenced by minimum temperature, maximum temperature, and precipitation (Chidumayo, 2001 ). Alterations in tree species composition or canopy structure will most probably have cascading effects on the understory vegetation (cf. Shirima et al., 2015), nutrient cycling, and water retention (cf. Guerrieri et al., 2021 and Liu et al., 2023 ). The adverse impacts of climate change on the functionality of the miombo ecosystem pose a significant threat to the livelihoods of those depending on miombo woodlands. To understand the effects of land use and climate changes on miombo woodland development for improved management, there is a need for accurate, complete, and concise information on the status and changes in forest cover, tree species composition, and canopy phenology. Thus, long-term monitoring of the miombo woodland is crucial for establishing baselines against which changes, such as land use alterations or climate change, can be assessed (cf. Lindenmayer et al., 2012 ). This includes detecting changes in the areal extension, quality, or structures, such as alterations in tree species composition, biomass, and canopy cover. Remote sensing, particularly applying optical sensors, is pivotal in monitoring canopy cover, providing a cost-effective and efficient alternative to ground monitoring, especially in challenging terrains (Petersen et al., 2018 ). Remote sensing methods can be categorized into wall-to-wall mapping and sample-based approaches, with wall-to-wall mapping being favored in forest cover monitoring for its ability to provide detailed spatial information. However, wall-to-wall mapping efforts may produce varying estimates for the same area and time, primarily because the estimation process relies on validating the model used to generate the estimates. These efforts also have limitations in providing measures of uncertainty and bias (Corona et al., 2015 ). The risk of misestimating canopy cover using optical sensors is even greater in dryland forests like miombo woodlands due to factors such as deciduousness, wildfires, variations in tree cover based on soil conditions (e.g., open and closed covers), and cloud cover during the wet season (Bodart et al., 2013 ; Verhegghen et al., 2022 ). Given these challenges, design-based sampling approaches are better for obtaining canopy cover estimates with known statistical precision and validating wall-to-wall mapping results (Corona et al., 2015 ; Sannier et al., 2014 ). Stratified random sampling and systematic sampling of Landsat imageries (30 m) have been applied to estimate forest cover and forest cover change in dry African ecoregions (Bodart et al., 2013 ; FAO, 2022 ). However, both approaches may not necessarily capture the variations within vegetation types, such as miombo woodlands in Tanzania. This highlights the need for a robust approach, like a spatially balanced sampling (Benedetti et al., 2017 ; Deville & Tille, 2004 ; Grafström & Schelin, 2014 ). A spatially balanced sample ensures that the distribution of auxiliary variables matches the sampling frame’s, allowing the random sample to accurately represent the forest conditions concerning the chosen auxiliary variables. This balanced selection of tracts enhances estimation precision, meaning that the estimates are closer to the true values than traditional sampling methods. The design is also flexible, allowing scalability and the inclusion of uncommon and common phenomena within the same framework through different sampling densities, multistage sampling, and the exclusion of tracts (Grafström & Matei, 2018 ). The potential of a balanced sampling design has been increasingly recognized for large-scale environmental monitoring and forest cover change assessment, as demonstrated in studies by Kermorvant et al. ( 2019 ) and Pagliarella et al. ( 2018 ). This study employed a multi-level sampling approach to estimate miombo woodland canopy cover in Tanzania. Specifically, we demonstrate the use of a two-stage sampling design combined with spatially balance sampling to provide a precise estimate of canopy cover within the miombo woodland of Western Tanzania. Materials and methods Study area We conducted the study in the Mlele and Tanganyika districts in the Katavi Region, as well as in the Sikonge and Tabora districts in the Tabora Region, all located in Western Tanzania (Table 1 , Fig. 1 ). The miombo woodlands in the districts experience unimodal rainfall pattern starting in November and ending in April of the following year. The dominant soil type is Cambisols. The climate and soil support both dry and wet miombo woodland vegetation types. The miombo canopy cover in the study area was estimated at 40,163 km 2 in 2009 during the National Forestry Resources Monitoring and Assessment (NAFORMA) (URT, 2015 ). Subsequent estimates indicate variations in tree cover for the study area, with values of 49,195 km 2 in 2018 from the Verhegghen et al. ( 2022 ) tree cover/non-tree cover dataset, 30,035 km 2 in 2020 from the ESA World cover dataset (Zanaga et al., 2021 ), and 41,710 km 2 in 2021 from the ESRI land cover dataset (Karra et al., 2021 ). Table 1 Description of districts comprising the study area District Land area (km 2 ) Location range Elevation range (masl) Climate ranges Miombo type Latitude Longitude Rainfall (mm) Temperature (℃) Tabora 1460 4° 40' 12''S − 5° 12' 47''S 32° 38' 20''E − 32° 59' 38''E 1123–1394 916–1008 23–24 Dry Sikonge 26260 5° 11' 42''S − 6° 58' 34''S 32° 01' 44''E − 34° 06' 36''E 1059–1632 648–1010 20–24 Dry Mlele 15558 5° 42' 47''S − 7° 16' 08''S 31° 25' 30''E − 32° 45' 25''E 832–1670 858–1054 20–25 Dry and Wet Tanganyika 16932 5° 07' 59''S − 6° 56' 13''S 29° 54' 11''E − 31° 11' 10''E 759–2034 942–1244 19–26 Wet Sampling design We utilized a two-stage sampling design that incorporated spatially balanced sampling to select 1 km x 1 km tracts in the first stage, followed by stratified random sampling that chose 30 m x 30 m plots in the second stage. The processes involved in the selection of tracts and plots are as follows: Selection of 1 km x 1 km tracts in the first stage We generated a sampling frame of 998,925 tracts of 1 km x 1 km that covered the entire country and prepared auxiliary variables (as listed in Table 2 ) that characterized miombo woodland distribution in Tanzania. The preparation of auxiliary variables involved collecting data from the specified sources and harmonizing their coordinate reference systems and spatial resolutions. The coordinate reference system of the tracts was also matched with those of auxiliary data so that they could overlay. Table 2 Auxiliary variables used in the balance sampling. Auxiliary variable name Justification/Explanation Source Data description Woodland cover It depicted the distribution of miombo woodland in Tanzania for the year 2019. The layer was generated by subtracting deforested areas between 2013 and 2019, as indicated in Hansen's loss year 2019, from the National Land Cover 2013. The processing was conducted using ArcMap software. The National land cover of 2013 and the Hansen dataset for 2019 (Loss year) (Hansen et al., 2023 ) Miombo woodland distribution in Tanzania at 30m resolution NDVI pattern It measures vegetation health and can be used to distinguish between different vegetation types. MODIS Q1 Vegetation indices (NDVI and EVI 16-Day L3 Global 250m) Vegetation indices for August 2021 Elevation It is one of the sources of heterogeneity in miombo woodlands (Mwakalukwa et al., 2014 ). JAXA DEM ( https://www.eorc.jaxa.jp/ALOS/en/aw3d30/data/index.htm ) Digital Elevation Model 30 m resolution Rainfall ● It is used to characterize miombo woodlands into dry and wet types (Jinga & Palagi, 2020 ; Lilleso et al., 2014 ). ● It is an environmental factor affecting the distribution of miombo tree species (Chidumayo, 2017 ; Jinga & Palagi, 2020 ). WorldClim precipitation data at 2.5 m (21 km 2 ) resolution 1970–2000. Monthly precipitation (mm). The data were summarized to obtain –total annual rainfall. Soil Moisture Index It is related to rainfall and is an environmental factor affecting the distribution of miombo tree species. SMOS L2 Science data 2500 m was obtained from the ESA site. ( https://smos-diss.eo.esa.int/smos/faq.html#Available%20data ) Soil Moisture (m3m-3) for August 2021 Soil Organic Carbon Stock It is a soil nutrient that determines miombo species composition (Mwakalukwa et al., 2014 ). FAO GLOSIS at 500 m resolution - Global Soil Organic Carbon Map (v1.5.0). ( http://54.229.242.119/GSOCmap ) Soil Organic Carbon in tonnes/ha Total nitrogen It is a soil nutrient that determines miombo species composition (Mwakalukwa et al., 2014 ). ISRIC World Soil Information at 250 m resolution. ( https://data.isric.org/geonetwork/srv/eng/catalog.search#/metadata/c5cb0f71-a627-40b9-8acb-5be42e4d5914 ) Total Nitrogen (N) content in mg/kg (ppm) spatially predicted for 0–30 cm depth. Total phosphorus It is one of the soil nutrients that determine miombo species composition (Shirima, 2015 ). ISRIC World Soil Information at 250 m resolution. ( https://data.isric.org/geonetwork/srv/eng/catalog.search#/metadata/f72f5698-4a3d-4af2-ae12-a4b1cf151cec ) Total Phosphorus (P) content in mg/kg (ppm) for 0–30 cm depth. We extracted values from each auxiliary layer for every tract. These extracted values were then used as inputs for the Balanced Sampling tool in R software (Grafström & Lisic, 2019 ), which sampled 1000 tracts across Tanzania, ensuring equal inclusion probabilities. Of the 998,925 tracts that covered Tanzania, 60,210 tracts covered the study area. Further, of the 1000 tracts sampled by the Balanced Sampling tool, 68 were located within our study area (Fig. 2 ). Selection of 30 m x 30 m plots in the second stage Each of the selected 68 tracts comprised an average of 1025 plots of 30 m x 30 m, categorized based on the percentage of woodland cover, as shown in Table 3 . The miombo woodland cover was determined by classifying Sentinel-2 images from 2020/2021 using the random forest method (Appendix 1). The definitions for these woodland cover classes were adopted from the United Nations Environment Program (UNEP), which classifies 'closed canopy' as having a canopy cover of more than 40% and 'open canopy' as having a canopy cover of 10–40%. The Tropical Ecosystem Environment Observations by Satellite (TREES) project also defines a 'dense canopy' with more than 70% canopy cover. (Tchatchou et al., 2015 ). We applied stratified random sampling to select a maximum of 40 plots from the categorized 1025 plots within each of the selected 68 tracts. Table 3 details the number of plots in the sampling frame and the number of plots selected for each canopy cover category in the sampling frame. Table 3 Number of plots and number of sampled plots per cover categories in selected tracts Cover categories Number of plots Number of sampled plots Closed miombo woodland (canopy cover > 70% ) 32442 1338 Open miombo woodland (40% ≤ canopy cover ≤ 70%) 6683 201 Very open miombo woodland (10% ≤ canopy cover < 40%) 7901 228 Non-miombo woodlands (canopy cover < 10%) 22690 923 Total 69716 2690 Preparation of reference datasets In this study, we generated classified maps of woodland and non-woodland for each tract by analyzing Planetscope imagery. These maps served as reference datasets for collecting canopy cover data per plot. The decision to utilize remote sensing data, specifically Planetscope imagery, was prompted by its cost-effectiveness compared to ground data collection. Additionally, remote sensing offers a standardized approach to measuring canopy cover, minimizing variations from different observers and collection times (seasons) associated with ground-based methods. In contrast to wall-to-wall mapping, image classification at the tract level is more accurate due to more intensive effort applied to smaller tracts (cf. Stehman et al. ( 2003 ). PlanetScope basemaps acquisition PlanetScope basemaps, Normalized Analytic, were acquired from the National Imaging and Mapping Agency (NICI) ( https://www.planet.com/basemaps/#/mode/selection/mosaic ). PlanetScope was selected for its high spatial resolution (3–5 m), making its images suitable for visual interpretation (Asrat et al., 2018 ), and its daily availability, enhancing the probability of obtaining monthly cloud-free images. Before selecting the monthly images, we analyzed the bands for each month between January 2022 and December 2022. The analysis showed November and December 2022 as the most suitable months for mapping miombo vegetation in the study area. This choice aligned with the timing of miombo species leaf bud development, which began in August before the October rains. In November, miombo trees in the lowlands exhibited full leaf canopies, and by December, the majority of miombo tree species in both lowlands and highlands displayed full leaf canopies. Additionally, in November and December, the presence of crops on farms and the occurrence of burned areas had minimal effects on spectral signals. Thus, we utilized the imagery from November and December 2022. Additionally, January 2023 images were included to increase the likelihood of obtaining cloud-free imagery. PlanetScope imageries processing Mapping miombo woodland from the collected planetScope images was done through Normalized Difference Vegetation Index (NDVI) values, followed by a visual verification step. NDVI layers were generated using the formula (1), and the miombo woodland class was separated by NDVI thresholds that ranged between 0.5 and 0.65. The overall classification accuracy of PlanetScope imagery within tracts was 0.94, accompanied by a Kappa statistic of 0.61 (Appendix 2). $$NDVI= \frac{NIR-Red}{NIR+Red} \left(1\right)$$ Where NIR = Near Infrared band and Red = Red band Data collection Data were collected in sampled 30 m x 30 m plots within each tract. The miombo woodland class, prepared for each tract as described above, was overlaid with the sampled plots. Subsequently, the canopy cover, representing the percentage of the miombo woodland in the plot, was determined in each plot. This implies that canopy cover was assessed as viewed from above by PlanetScope (cf. Westfall & Morin, 2013 ). This process is also known as diffuse canopy cover measurements, wherein overlapping crowns from neighboring trees are combined. Data analysis Canopy Cover Estimation The Horvitz–Thompson (H.T.) estimator was chosen for its compatibility with the Spatial Balanced Design (Grafström & Tillé, 2013 ). Using equations (1)–(3) in Appendix (3), the estimator determined the total area of miombo woodland with closed, open, and very open canopy covers based on the number of plots in each canopy cover category, as identified from the classified PlanetScope images. Subsequently, equations (4)–(8) of the H.T. estimator (in Appendix 3) were used to estimate the canopy cover within the miombo woodland for closed, open, and very open canopy covers, utilizing the collected data on canopy cover. The standard error was computed as the square root of the variance estimated from the equations, and the relative standard error (RSE) was calculated as the ratio of the standard error to the estimated mean or total. Results and Discussion Results Canopy cover estimates The findings reveal that miombo woodland coverage within our study area was 37,359 ± 4,618 km², representing approximately 54–70% of the study area (Table 4 ). Among different miombo woodland classes, the closed canopy cover type contributed the largest portion, covering an area of 29,546 ± 4,382 km², constituting about 42–56% of the study area. The very open canopy cover type followed closely, with an area of 4,094 ± 1,127 km², making up approximately 5–9% of the study area. The smallest contribution was from the open canopy cover type, with an area of 3,719 ± 849 km², accounting for about 5–8% of the study area. Table 4 Estimates of miombo woodland area (in square kilometers) for 2022/2023 Canopy cover class Area (km 2 ) ME (± 95% confidence interval) RSE (%) The miombo woodland area with closed canopy cover (cover% > 70%) 29546 4382 7.57 The miombo woodland area with open canopy cover (40% ≤ cover% ≤ 70%) 3719 849 11.64 The miombo woodland area with very open canopy cover (10% ≤ cover% < 40%) 4094 1127 14.04 The miombo woodland area (≥ 10%) 37359 4618 6.31 Furthermore, the findings indicate that the canopy cover in the miombo woodland area was 55% ± 5%. Consequently, the miombo woodland in the study area predominantly fell within the open canopy cover class. Regarding miombo woodland classes, the canopy cover within the areas classified as closed, open, and very open cover classes are presented in Table 5 below. Table 5 Estimates of canopy cover in the miombo woodland area Canopy cover class Canopy cover (%) ME (± 95% confidence interval) RSE (%) The canopy cover in the miombo woodland area with closed canopy cover (cover% > 70%) 84 7 4.50 The canopy cover in the miombo woodland area with open canopy cover (40% ≤ cover% ≤ 70%) 45 6 6.34 The canopy cover in the miombo woodland area with very open canopy cover (10% ≤ cover% < 40%) 19 3 6.91 The canopy cover in the miombo woodland area (≥ 10%) 55 5 4.63 Discussion In this paper, we demonstrate the design of a monitoring scheme for miombo woodland in Tanzania, producing accurate and precise estimates of its areal extent and canopy cover in the study area. Notably, all estimates displayed Relative Standard Errors (RSEs) well below 25%, indicating their reliability, as Sileshi ( 2015 ) suggested. The results reveal that the miombo woodland in the study area covers 37,359 ± 4,618 km 2 , with an average canopy cover of 55% ± 5%. The closed miombo woodland, with an average canopy cover of 84% ± 7%, comprised 79% of the miombo woodland area. Open miombo woodland, with an average canopy cover of 45% ± 6%, and very open miombo woodland, with an average canopy cover of 19% ± 3%, comprised 10% and 11% of the miombo woodland area, respectively. This could be attributed to the fact that 74% of the miombo woodland in the study area is within protected areas and influenced by the presence of wet miombo woodlands. A closed canopy characterizes wet miombo woodland due to a positive correlation between rainfall and miombo growth rates (Frost, 1996 ). However, the results regarding canopy cover do not align with the estimates provided by the Moderate Resolution Imaging Spectroradiometer vegetation continuous fields (MODIS VCF) tree cover product for the year 2022 (cf. DiMiceli et al., 2017 ). According to the MODIS VCF data, very open miombo woodland, open miombo woodland, and closed miombo woodland covers an area of 40,680 km 2 , 2,557 km 2 , and 6 km 2 , respectively, totaling 43,243 km 2 of the miombo woodland area. This suggests a dominance of very open miombo woodlands, according to MODIS VCF. Contrastingly, the findings of Adzhar et al. ( 2021 ) indicate that MODIS VCF consistently underestimates woody cover in tropical savannas. The observed discrepancy raises questions about the accuracy of MODIS VCF estimates, particularly in characterizing miombo woodland cover in the studied area. The area of miombo woodland found in this study align with the estimates from the ESRI land cover dataset that revealed the trees covered 41,710 km 2 in 2021 (Karra et al., 2021 ). Other recent land cover assessment estimates revealed the trees covered 49,195 km 2 in 2018 (Verhegghen et al., 2022 ) and 30,035 km 2 in 2020 (Zanaga et al., 2021 ). However, the estimates of Karra et al. ( 2021 ), Verhegghen et al. ( 2022 ), and Zanaga et al. ( 2021 ) from wall-to-wall mapping lack standard errors, contributing to difficulty in assessing the reliability and precision of their estimates for better comparison with our study estimates. Wall-to-wall mapping, which is suitable for providing spatially explicit information, contains errors arising from data, classification algorithms, and analysts that affect the accuracy of land cover estimations. To address these uncertainties, evaluations of wall-to-wall mapping often involve comparing the produced land cover with ground truth data using a confusion matrix. In order to offer an unbiased assessment of wall-to-wall mapping, a design-based statistical sampling is employed to estimate land cover with confidence levels (Corona, 2010 ; Corona et al., 2015 ). This approach helps mitigate errors and provides a more robust estimation framework (cf. Sannier et al. 2014 ). Additionally, a sample-based approach to land cover estimation proves to be cost-effective and time-efficient, as efforts are concentrated in a smaller representative area. Due to the estimates of miombo woodland obtained in this study and the advantages of statistical sampling, our results support the idea that wall-to-wall mapping should be integrated with sample-based assessment (cf. Maniatis et al. 2021 and Gallaun et al. 2015 ). This integration will help reduce uncertainties resulting from the challenges associated with dryland canopy cover monitoring, as Bodart et al. ( 2013 ) observed. Furthermore, our results indicate an improvement in the reliability of estimates with increasing canopy cover, as indicated by the decreasing Relative Standard Error (RSE) from very open to open and closed canopy cover classes. This improvement is linked to sample size, where a closed woodland class was the most common, and the homogeneity of measured canopy cover within the class. Despite a smaller sample size within the open woodland class compared to the very open woodland class, the RSE within the open woodland class was lower, reflecting reduced variability in the measured canopy cover within this class. Corona et al. ( 2015 ) similarly found that the greatest sampling effort resulted in a low Relative Standard Error (RSE), especially with many primary sampling units (n) at the first stage and few secondary sampling units (m) at the second stage, for a specific sample size (m x n) in the two-stage sampling strategy. Therefore, increasing the sample size, primarily through increasing the number of tracts in the first stage, will likely enhance the accuracy of estimates derived from the sampling strategy. The inventory design used in this study is based on the inventory design used in the national monitoring program NILS – National Inventories of Landscapes in Sweden (Allard et al. 2023 , Brown et al. 2023 , Ranlund et al. 2023 ; https://www.slu.se/en/Collaborative-Centres-and-Projects/nils/ ). The methodology involves among other things a multi-stage sampling approach. In the first stage, we utilize spatially balanced sampling to select tracts, and in the second stage, we utilize stratified random sampling to select plots within the tracts. This two-stage sampling approach is effective since only 0.1% of the tracts and 4% of the plots need to be inventoried. However, the tracts and plots that did not need to be inventoried in the planetScope in the second stage were included, contributing to the large sample size and high precision of the estimates. These were tracts and plots lacking the focal phenomena (in this case, miombo woodland). It was possible to include tracks and plots lacking focal phenomena, as the study's second stage involved an ocular inventory of selected plots. However, for costly and time-consuming field inventories, the focus can be on tracts and plots with the potential miombo woodland or with the potential miombo canopy cover class to ensure cost-effectiveness. Further, the design allows for different sampling methods in the second stage and between primary sampling units for the selection of secondary sampling units within them. The choice of methods is guided by the characteristics of the population and research objectives, which also might result in varying sample sizes of secondary sampling units between primary sampling units. This inherent flexibility positions the design as a potential model for an effective national monitoring program, seamlessly integrating remote sensing and field inventories. Conclusions Long-term monitoring of miombo woodlands is crucial to enhance their management, given the growing pressure from unsustainable harvesting and the impacts of climate and land use changes. This study established the reliability and precision of a monitoring design employing a two-stage sampling strategy that utilizes spatially balanced sampling in the first stage. Results from this approach revealed that miombo woodlands in Tabora, Sikonge, Mlele, and Tanganyika districts collectively cover 37,359 ± 4,618 km², with an average canopy cover of 55% ± 5%. Moreover, closed miombo woodland, with an average canopy cover of 84% ± 7%, dominates the landscape, covering an area of 29,546 ± 4,382 km². These results had relative standard errors consistently below 25%. These findings on miombo woodlands' canopy cover provide a baseline for monitoring miombo types and studying various ecological processes. The design exhibits effectiveness by enabling focused sampling and demonstrates flexibility by allowing the integration of varying sampling methods in the second stage. Furthermore, the proposed sampling strategy could potentially serve as a valuable tool for validating wall-to-wall mapping efforts and as an effective model for a national canopy monitoring program. Declarations Interests: The authors have no competing interests to declare that are relevant to the content of this article. All authors have read, understood, and have complied as applicable with the statement on "Ethical responsibilities of Authors' as found in the Instructions for Authors. Funding: This work was funded by Sida under Grant Agreement No 13394. Author Contribution Siwa Ernest Nkya, Deo Dominick Shirima, Henrik Hedenas, and August Temu contributed to the design and implementation of the research. Prior to writing the manuscript, Siwa Ernest Nkya and Henrik Hedenas analyzed the data. In the manuscript, Siwa Ernest Nkya and Henrik Hedenas wrote the introduction, results, and discussion sections; Siwa Ernest Nkya and Deo Dominick Shirima wrote the methodology section, and Siwa Ernest Nkya and August Temu wrote the abstract and conclusion sections. All authors revised the manuscript and approved its submission for publication to Environment Monitoring and Assessment. Data availability: The data supporting the findings of this study are available within the paper and its supplementary information files. References Abdallah, J., & Monela, G.G. (2007). Overview of Miombo woodlands in Tanzania. Proceedings of the First MITMIOMBO Project Work- Shop , 50 (February), 9–23. https://jukuri.luke.fi/bitstream/handle/10024/535965/mwp050.pdf?isAllowed=y&sequence=1 Adzhar, R., Kelley, D., Dong, N., Torello Raventos, M., Veenendaal, E., Feldpausch, T., Philips, O., Lewis, S., Sonké, B., Taedoumg, H., Schwantes Marimon, B., Domingues, T., Arroyo, L., Djagbletey, G., Saiz, G., & Gerard, F. (2021). Assessing MODIS Vegetation Continuous Fields tree cover product (collection 6): performance and applicability in tropical forests and savannas. Biogeosciences Discussions, February, 1–20. https://doi.org/10.5194/bg-2020-460 Allard, A., Brown, A., Hurford, C., Isendahl, C., Hilpold, A., Pappeiner, U., Strolb J. & Hedenås, H. 2023. Chapter 5. Data collected in situ, unique details or integrated components of monitoring schemes. In. Allard, A., Keskitalo, C.H. & Brown A. (reds). Monitoring Biodiversity Combining Environmental and Social Data. Routledge. Asrat, Z., Taddese, H., Ørka, H. O., Gobakken, T., & Næsset, E. (2018). Estimation of Forest Area and Canopy Cover Based on Visual Interpretation of Satellite Images in Ethiopia. Land , 7 (92), 1–17. https://doi.org/10.3390/land7030092 Benedetti, R., Piersimoni, F., & Postiglione, P. (2017). Alternative and complementary approaches to spatially balanced samples. Metron , 75 (3), 249–264. https://doi.org/10.1007/s40300-017-0123-1 Bodart, C., Brink, A. B., Donnay, F., Lupi, A., Mayaux, P., & Achard, F. (2013). Continental estimates of forest cover and forest cover changes in the dry ecosystems of Africa between 1990 and 2000. Journal of Biogeography , 40 (6), 1036–1047. https://doi.org/10.1111/jbi.12084 Brown, A., Hedenås, H., Holm, E., Lind, T., Richards, A.E., Prober, S.M. & Schmidt B. 2023. Chapter 15. Designing and Adapting Biodiversity Monitoring Schemes. In. Allard, A., Keskitalo, C.H. & Brown A. (reds). Monitoring Biodiversity Combining Environmental and Social Data. Routledge. Chidumayo, E. N. (2001). Climate and Phenology of Savanna Vegetation in Southern Africa. Journal of Vegetation Science, 12(3), 347. https://doi.org/10.2307/3236848 Chidumayo, E. N. (2017). Biotic interactions, climate and disturbance underlie the distribution of two Julbernardia tree species in miombo woodlands of Africa . 1–11. https://doi.org/10.1017/S0266467416000584 Chidumayo, E. N. (2019). Management implications of tree growth patterns in miombo woodlands of Zambia. Forest Ecology and Management, 436(October 2018), 105–116. https://doi.org/10.1016/j.foreco.2019.01.018 Corbera, E., Estrada, M., & Brown, K. (2010). Reducing greenhouse gas emissions from deforestation and forest degradation in developing countries: Revisiting the assumptions. Climatic Change , 100 (3), 355–388. https://doi.org/10.1007/s10584-009-9773-1 Corona, P. (2010). Integration of forest mapping and inventory to support forest management. IForest - Biogeosciences and Forestry , 59–64. https://doi.org/10.3832/ifor0531-003 Corona, P., Fattorini, L., & Pagliarella, M. C. (2015). Sampling strategies for estimating forest cover from remote sensing-based two-stage inventories. Forest Ecosystems , 2 (1). https://doi.org/10.1186/s40663-015-0042-7 Deville, B. J., & Tille, Y. (2004). Efficient balanced sampling : The cube method . 893–912. Dewees, P. ., Campbell, B. ., Katerere, Y., Sitoe, A., Cunningham, A. ., Angelsen, A., & Wunder, S. (2011). Managing the Miombo Woodlands of Southern Africa: Policies, incentives, and options for the rural poor . Program on Forests (PROFOR). DiMiceli, C. ., Carroll, M. ., Sohlberg, R. ., Huang, C., Hansen, M. C., & Townshend, J. R. G. (2017). Annual global automated MODIS vegetation continuous fields (MOD44B) at 250 m spatial resolution for data years beginning day 65, 2000-2014, collection 5 percent tree cover, version 6 (p. 12). University of Maryland, College Park, MD. https://lpdaac.usgs.gov/documents/1494/MOD44B_User_Guide_V61.pdf FAO. (2022). FRA 2020 Remote Sensing Survey. In FAO Forestry Paper No. 186 . https://doi.org/10.4060/cb9970en Frost, P. (1996). The ecology of Miombo woodlands. In B. Campbell (Ed.), The Miombo in Transition: Woodlands and Welfare in Africa (Issue June, pp. 11–57). Center for International Forestry Research (CIFOR). http://books.google.com/books?hl=nl&lr=&id=rpildJJVdU4C&pgis=1 Gallaun, H., Steinegger, M., Wack, R., Schardt, M., Kornberger, B., & Schmitt, U. (2015). Remote sensing based two-stage sampling for accuracy assessment and area estimation of land cover changes. Remote Sensing , 7 (9), 11992–12008. https://doi.org/10.3390/rs70911992 Grafström, A., & Lisic, J. (2019). Package ‘BalancedSampling .’ CRAN Reference Manual , 25. https://cran.r-project.org/web/packages/BalancedSampling/BalancedSampling.pdf Grafström, A., & Matei, A. (2018). Spatially Balanced Sampling of Continuous Populations. Scandinavian Journal of Statistics , 45 (3), 792–805. https://doi.org/10.1111/sjos.12322 Grafström, A., & Schelin, L. (2014). How to Select Representative Samples. Scandinavian Journal of Statistics , 41 (2), 277–290. https://doi.org/10.1111/sjos.12016 Grafström, A., & Tillé, Y. (2013). Doubly balanced spatial sampling with spreading and restitution of auxiliary totals. Environmetrics , 24 (2), 120–131. https://doi.org/10.1002/env.2194 Guerrieri, R., Templer, P., & Magnani, F. (2021). Canopy Exchange and Modification of Nitrogen Fluxes in Forest Ecosystems. Current Forestry Reports, 7, 115–137. https://doi.org/10.1007/s40725-021-00141-y Hansen, M. C., Potapov, P. V., Moore, R., Hancher, M., Turubanova, S. ., Tyukavina, A., Thau, D., Stehman, S. V., Goetz, S. J., Loveland, T. R., Kommareddy, A., Egorov, A., Chini, L., Justice, C. O., & Townshend, J. R. G. (2023). High-Resolution Global Maps of 21st-Century Forest Cover Change. Science 342 , 850 (53). http://earthenginepartners.appspot.com/science-2013-global-forest Jinga, P., & Palagi, J. (2020). Dry and wet miombo woodlands of south-central Africa respond differently to climate change. Environmental Monitoring and Assessment , 192:372 , 1–16. https://doi.org/https://doi.org/10.1007/s10661-020-08342-x John, E., Bunting, P., Hardy, A., Roberts, O., Giliba, R., & Silayo, D. S. (2020). Modelling the impact of climate change on Tanzanian forests. Diversity and Distributions , 26 (12), 1663–1686. https://doi.org/10.1111/ddi.13152 Karra, K., Kontgis, C., Statman-Weil, Z., Mazzariello, J., Mathis, M., & Brumby, S. (2021). Global land use/land cover with Sentinel-2 and deep learning. IGARSS 2021 IEEE International Geoscience and Remote Sensing Symposium , april . Kermorvant, C., D’Amico, F., Bru, N., Caill-Milly, N., & Robertson, B. (2019). Spatially balanced sampling designs for environmental surveys. Environmental Monitoring and Assessment , 191 (8). https://doi.org/10.1007/s10661-019-7666-y Lilleso, J.-P. ., van Breugel, P., Kindt, R., Mbago, F., Moshi, H. ., Ndangalasi, H. ., Uronu, L. O. ., Jamnadass, R., & Graudal, L. (2014). Potential Natural Vegetation of Eastern Africa ( Ethiopia , Kenya , Malawi , Rwanda , Tanzania , Uganda and Zambia ). Volume 10. Atlas and Tree Species Composition for Tanzania (Vol. 10, Issue 62). Department of Geosciences and Natural Resource Management, University of Copenhagen. https://ign.ku.dk/english/research/forest-nature-biomass/tropical-trees-landscapes/software/vecea/series/filer/Tanzania_Volume_10_web.pdf Lindenmayer, D. B., Likens, G. E., Andersen, A., Bowman, D., Bull, C. M., Burns, E., Dickman, C. R., Hoffmann, A. A., Keith, D. A., Liddell, M. J., Lowe, A. J., Metcalfe, D. J., Phinn, S. R., Russell-Smith, J., Thurgate, N., & Wardle, G. M. (2012). Value of long-term ecological studies. Austral Ecology , 37 (7), 745–757. https://doi.org/10.1111/j.1442-9993.2011.02351.x Liu, Y., Shi, W., Tao, F., Shi, X., & Fu, B. (2023). A global synthesis of multi‐factors affecting water storage capacity in forest canopy, litter and soil layers. Geophysical Research Letters, 50. Lupala, Z. J., Lusambo, L. P., & Ngaga, Y. M. (2014). Management, Growth, and Carbon Storage in Miombo Woodlands of Tanzania. International Journal of Forestry Research , 2014 , 1–11. https://doi.org/10.1155/2014/629317 Lusambo, L. P. (2009). Economics on the Household energy in miombo woodlands of Eastern and Southern Tanzania . University of Bangor. Maniatis, D., Dionisio, D., Guarnieri, L., Marchi, G., Mollicone, D., Díaz, A. S. P., & Morales, C. (2021). Toward a more representative monitoring of land-use and land-cover dynamics: The use of a sample-based assessment through augmented visual interpretation using open foris collect earth. Remote Sensing , 13 (21). https://doi.org/10.3390/rs13214197 Mittermeier, R. A., Mittermeier, C. G., Brooks, T. M., Pilgrim, J. D., Konstant, W. R., Fonseca, G. A. B., & Kormos, C. (2003). Wilderness and biodiversity conservation. Proceedings of National Academy of Sciences, 100(18), 10309–10313. https://doi.org/10.1073/pnas.1732458100 Mwakalukwa, E. E., Meilby, H., & Treue, T. (2014). Floristic Composition, Structure, and Species Associations of Dry Miombo Woodland in Tanzania. ISRN Biodiversity , 2014 , 1–15. https://doi.org/10.1155/2014/153278 Pagliarella, M. C., Corona, P., & Fattorini, L. (2018). Spatially-balanced sampling versus unbalanced stratified sampling for assessing forest change: evidences in favour of spatial balance. Environmental and Ecological Statistics , 25 (1), 111–123. https://doi.org/10.1007/s10651-017-0378-y Parry, M. ., Canziani, O. ., Palutikof, J. ., & Co-authors. (2007). Technical summary. In M. . Parry, O. . Canziani, J. . Palutikof, P. . van der Linden, & E. Hanson, C.E (Eds.), Climate Change 2007: Impacts, Adaptation and Vulnerability. Contribution of Working Group II to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change (pp. 23–78). Cambridge University Press. https://doi.org/10.1016/B978-008044910-4.00250-9 Petersen, R., Davis, C., Herold, M., & Sy, V. D. E. (2018). Tropical forest monitoring: Exploring the gaps between what is required and what is possible for REDD+ and other initiatives (Issue June, pp. 1–12). World Resources Institute. wri.org/ending-tropicaldeforestation Pfeifer, M., Gonsamo, A., Woodgate, W., Cayuela, L., Marshall, A. R., Ledo, A., Paine, T. C. E., Marchant, R., Burt, A., Calders, K., Courtney-Mustaphi, C., Cuni-Sanchez, A., Deere, N. J., Denu, D., de Tanago, J. G., Hayward, R., Lau, A., Macía, M. J., Olivier, P. I., … Platts, P. J. (2018). Tropical forest canopies and their relationships with climate and disturbance: results from a global dataset of consistent field-based measurements. Forest Ecosystems , 5 (1), 1–14. https://doi.org/10.1186/s40663-017-0118-7 Ranlund, Å., Grafström, A., Brown, A., Hedenås, H. & Levin, G. 2023. Chapter 4. Designing monitoring systems. In. Allard, A., Keskitalo, C.H. & Brown A. (reds). Monitoring Biodiversity Combining Environmental and Social Data. Routledge. Sannier, C., McRoberts, R. E., Fichet, L. V., & Makaga, E. M. K. (2014). Using the regression estimator with landsat data to estimate proportion forest cover and net proportion deforestation in gabon. Remote Sensing of Environment , 151 , 138–148. https://doi.org/10.1016/j.rse.2013.09.015 Shirima, D. D. (2015). PhD Thesis: Forests and woodlands of Tanzania: interactions between woody plant structure, diversity, carbon stocks and soil nutrient heterogeneity [Norwegian University of Life Sciences]. https://nmbu.brage.unit.no/nmbu-xmlui/handle/11250/2579210 Sileshi, G. W. (2015). The relative standard error as an easy index for checking the reliability of regression coefficients. Researchgate , Discussion paper , August , 1–25. https://doi.org/10.13140/RG.2.1.2123.6968 Stehman, S. V., Sohl, T. L., & Loveland, T. R. (2003). Statistical sampling to characterize recent United States land-cover change. Remote Sensing of Environment, 86(4), 517–529. https://doi.org/10.1016/S0034-4257(03)00129-9 Syampungani, S., Chirwa, P. W., Akinnifesi, F. K., Sileshi, G., & Ajayi, O. C. (2009). The Miombo woodlands at the cross roads: Potential threats, sustainable livelihoods, policy gaps and challenges. Natural Resources Forum , 33 (2), 150–159. https://doi.org/10.1111/j.1477-8947.2009.01218.x Tchatchou, B., Sonwa, D. ., Ifo, S., & Tiani, A. . (2015). Deforestation and forest degradation in the Congo Basin: State of knowledge , current causes and perspectives (No. 144). https://doi.org/10.17528/cifor/005894 URT. (2015). National Forest Resources Monitoring and Assessment of Tanzania Mainland: Main Results . United Republic of Tanzania, Ministry of Natural Resources and Tourism. Dar Es Salaam. Verhegghen, A., Kuzelova, K., Syrris, V., & Eva, H. (2022). Mapping Canopy Cover in African Dry Forests from the Combined Use of Sentinel-1 and Sentinel-2 Data : Application to Tanzania for the Year 2018. Remote Sensing , 14 (1522), 1–21. https://doi.org/https://doi.org/10.3390/rs14061522 Westfall, J. A., & Morin, R. S. (2013). A cover-based method to assess forest characteristics using inventory data and GIS. Forest Ecology and Management , 298 , 93–100. https://doi.org/10.1016/j.foreco.2013.02.036 Zanaga, D., Van De Kerchove, R., De Keersmaecker, W., Souverijns, N., Brockmann, C., Quast, R., Wevers, J., Grosu, A., Paccini, A., Vergnaud, S., Cartus, O., Santoro, M., Fritz, S., Georgieva, I., Lesiv, M., Carter, S., Herold, M., Linlin, L., Tsendbazar, N., … Arino, O. (2021). ESA WorldCover 10 m 2020 v100. Meteosat Second Generation Evapotranspiration (MET) . Additional Declarations No competing interests reported. Supplementary Files DataavailablemiombocanopycoverpaperSNDSHHAT.zip Appendices.docx 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-3880805","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":274576033,"identity":"c53ef67f-899b-40d9-b17c-940d7d930f48","order_by":0,"name":"Siwa E. Nkya","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA00lEQVRIiWNgGAWjYDACHhBhwMDAD6ITCkjRItkA0mJAtBaQrgNQvQSBOc8Zw48/Cg7LG59fnfjhgQGDPL/YAfxaLHt7jKV5DA4bbrvxdrME0GGGM2cn4NdicJ7HQJrB4DbjthtnN4C0JBjcJqzF+OcPg9v2m2ec3fyDOC1ne8wkeAxuJ27g791GpC1njpVZ8xj8T55xg3ebRYKBBBF+OZO8+eaPP2m2/f1ngYwKG3l+aQJaGBg4oHEhAVYpQUg5CLA/gND8B4hRPQpGwSgYBSMRAADlPUaVT1Um5wAAAABJRU5ErkJggg==","orcid":"","institution":"Regional Research School in Forest Sciences","correspondingAuthor":true,"prefix":"","firstName":"Siwa","middleName":"E.","lastName":"Nkya","suffix":""},{"id":274576034,"identity":"599fa8fb-fa17-4161-869f-229e841f7d9b","order_by":1,"name":"Deo D. Shirima","email":"","orcid":"","institution":"Sokoine University of Agriculture","correspondingAuthor":false,"prefix":"","firstName":"Deo","middleName":"D.","lastName":"Shirima","suffix":""},{"id":274576035,"identity":"48be4b02-71b5-4166-b8ef-f8d3263d87a3","order_by":2,"name":"Henrik Hedenas","email":"","orcid":"","institution":"Swedish University of Agricultural Sciences","correspondingAuthor":false,"prefix":"","firstName":"Henrik","middleName":"","lastName":"Hedenas","suffix":""},{"id":274576036,"identity":"e4d8b550-04d6-443f-909b-4b363e3272e0","order_by":3,"name":"August B. Temu","email":"","orcid":"","institution":"","correspondingAuthor":false,"prefix":"","firstName":"August","middleName":"B.","lastName":"Temu","suffix":""}],"badges":[],"createdAt":"2024-01-20 06:14:11","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-3880805/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-3880805/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":51715633,"identity":"774fff1d-b655-479e-9201-5fb1ed72f6e8","added_by":"auto","created_at":"2024-02-27 20:58:54","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":87783,"visible":true,"origin":"","legend":"\u003cp\u003eMap of the\u003cstrong\u003e \u003c/strong\u003estudy area\u003c/p\u003e","description":"","filename":"floatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-3880805/v1/40097c3502c707672e401e81.png"},{"id":51715634,"identity":"68f749e6-f404-4ef9-b165-5ee004c82155","added_by":"auto","created_at":"2024-02-27 20:58:54","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":632455,"visible":true,"origin":"","legend":"\u003cp\u003eAn illustration of the stages involved in the sampling design and its application in the inventory of miombo woodland. a) The sampling frame (grid 1 x 1 km2) and tracts were randomly selected from this frame using a balanced sample in the first stage, covering the entire Tanzania. b) Selected tracts within the study area. c) Plots within the tracts were automatically classified based on canopy cover and used as sampling classes. d) Stratified random sampling was employed in the second stage to select plots for assessing canopy cover.\u003c/p\u003e","description":"","filename":"floatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-3880805/v1/ea5cda78e5e05344d0015955.png"},{"id":55309437,"identity":"4962edf0-7548-4aab-a44b-00d71ff26bb0","added_by":"auto","created_at":"2024-04-25 14:15:47","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":1760310,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-3880805/v1/7c0b060a-7110-458d-8e33-158afa125230.pdf"},{"id":51715675,"identity":"53df0080-a9d1-4f05-9d23-8173f61834e8","added_by":"auto","created_at":"2024-02-27 20:59:13","extension":"zip","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":286415614,"visible":true,"origin":"","legend":"","description":"","filename":"DataavailablemiombocanopycoverpaperSNDSHHAT.zip","url":"https://assets-eu.researchsquare.com/files/rs-3880805/v1/e29b058810081a5983a04dd0.zip"},{"id":51715632,"identity":"020f32cc-5570-462c-91fa-4bcbf9228457","added_by":"auto","created_at":"2024-02-27 20:58:54","extension":"docx","order_by":2,"title":"","display":"","copyAsset":false,"role":"supplement","size":743732,"visible":true,"origin":"","legend":"","description":"","filename":"Appendices.docx","url":"https://assets-eu.researchsquare.com/files/rs-3880805/v1/4f925984398a31e72611586e.docx"}],"financialInterests":"No competing interests reported.","formattedTitle":"Estimating canopy cover using a Spatially Balanced Sampling approach: A case study of miombo woodlands in Western Tanzania","fulltext":[{"header":"Introduction","content":"\u003cp\u003eMiombo refers to woodlands dominated by Brachystegia, Julbernadia and/or isoberlinia genera from the legume family - Fabaceae, subfamily Caesalpinioideae. These woodlands are extensive tropical African formations covering 2.7\u0026nbsp;million km\u003csup\u003e2\u003c/sup\u003e within Angola, Malawi, Mozambique, Tanzania, Zimbabwe, Zambia, and the Democratic Republic of Congo (Frost, \u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e1996\u003c/span\u003e). In Tanzania, these woodlands cover 447263 km\u003csup\u003e2\u003c/sup\u003e, constituting about 93% of the total forested area (URT, \u003cspan citationid=\"CR52\" class=\"CitationRef\"\u003e2015\u003c/span\u003e). Miombo woodlands can be categorized into dry miombo, in areas receiving annual rainfall of less than 1000 mm, and wet miombo, in areas receiving yearly rainfall equal to or above 1000 mm (Frost, \u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e1996\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eDue to their extensiveness, miombo woodlands play a major role in poverty alleviation by providing household energy, timber, and food like mushrooms, fruits, and honey (Abdallah \u0026amp; Monela, \u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e2007\u003c/span\u003e; Dewees et al., \u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e2011\u003c/span\u003e; Lusambo, \u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e2009\u003c/span\u003e; Syampungani et al., \u003cspan citationid=\"CR50\" class=\"CitationRef\"\u003e2009\u003c/span\u003e). Miombo and mopane (another dry woodland in Southern Africa) harbors 8,500 plant species, of which 4,600 are endemic (Mittermeier et al., \u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e2003\u003c/span\u003e). In recent years, their role in carbon sequestration has also been recognized (Lupala et al., \u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e2014\u003c/span\u003e). The increasing demand for forest products and the expansion of cropland due to population increases result in woodland loss, thereby threatening the sustainable provision of these goods and services (Lusambo, \u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e2009\u003c/span\u003e; Syampungani et al., \u003cspan citationid=\"CR50\" class=\"CitationRef\"\u003e2009\u003c/span\u003e; URT, \u003cspan citationid=\"CR52\" class=\"CitationRef\"\u003e2015\u003c/span\u003e). Moreover, forest loss contributes to global warming and climate change (Corbera et al., \u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e2010\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eJohn et al. (\u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e2020\u003c/span\u003e) predicted that climate change will likely result in the loss of 5% of miombo woodland suitable habitat by 2085 due to a drier climate. However, the most significant impact might not be the net loss but rather the projected conversion of wet to dry miombo woodland. Jinga and Palagi (\u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e2020\u003c/span\u003e) predicted a decline in the areal extent of wet miombo by up to 41.6% by 2070, while the areal extent of dry miombo might increase by 22.7% over the same period. Thus, tree species adapted to drier conditions will be increased, replacing those miombo species thriving in wetter conditions.\u003c/p\u003e \u003cp\u003eSince dry miombo woodlands exhibit lower tree growth rates than wet miombo (Chidumayo \u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e2019\u003c/span\u003e), carbon sequestration may be further reduced. Moreover, even if there is no shift in tree species composition, climate change could impact the woodland canopy. Canopy structures are adapted to the maximum water deficits in the tropics (Pfeifer et al., \u003cspan citationid=\"CR44\" class=\"CitationRef\"\u003e2018\u003c/span\u003e), and the canopy phenology in the miombo is influenced by minimum temperature, maximum temperature, and precipitation (Chidumayo, \u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e2001\u003c/span\u003e). Alterations in tree species composition or canopy structure will most probably have cascading effects on the understory vegetation (cf. Shirima et al., 2015), nutrient cycling, and water retention (cf. Guerrieri et al., \u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e2021\u003c/span\u003e and Liu et al., \u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). The adverse impacts of climate change on the functionality of the miombo ecosystem pose a significant threat to the livelihoods of those depending on miombo woodlands.\u003c/p\u003e \u003cp\u003eTo understand the effects of land use and climate changes on miombo woodland development for improved management, there is a need for accurate, complete, and concise information on the status and changes in forest cover, tree species composition, and canopy phenology. Thus, long-term monitoring of the miombo woodland is crucial for establishing baselines against which changes, such as land use alterations or climate change, can be assessed (cf. Lindenmayer et al., \u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e2012\u003c/span\u003e). This includes detecting changes in the areal extension, quality, or structures, such as alterations in tree species composition, biomass, and canopy cover. Remote sensing, particularly applying optical sensors, is pivotal in monitoring canopy cover, providing a cost-effective and efficient alternative to ground monitoring, especially in challenging terrains (Petersen et al., \u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e2018\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eRemote sensing methods can be categorized into wall-to-wall mapping and sample-based approaches, with wall-to-wall mapping being favored in forest cover monitoring for its ability to provide detailed spatial information. However, wall-to-wall mapping efforts may produce varying estimates for the same area and time, primarily because the estimation process relies on validating the model used to generate the estimates. These efforts also have limitations in providing measures of uncertainty and bias (Corona et al., \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e2015\u003c/span\u003e). The risk of misestimating canopy cover using optical sensors is even greater in dryland forests like miombo woodlands due to factors such as deciduousness, wildfires, variations in tree cover based on soil conditions (e.g., open and closed covers), and cloud cover during the wet season (Bodart et al., \u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e2013\u003c/span\u003e; Verhegghen et al., \u003cspan citationid=\"CR53\" class=\"CitationRef\"\u003e2022\u003c/span\u003e). Given these challenges, design-based sampling approaches are better for obtaining canopy cover estimates with known statistical precision and validating wall-to-wall mapping results (Corona et al., \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e2015\u003c/span\u003e; Sannier et al., \u003cspan citationid=\"CR46\" class=\"CitationRef\"\u003e2014\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eStratified random sampling and systematic sampling of Landsat imageries (30 m) have been applied to estimate forest cover and forest cover change in dry African ecoregions (Bodart et al., \u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e2013\u003c/span\u003e; FAO, \u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e2022\u003c/span\u003e). However, both approaches may not necessarily capture the variations within vegetation types, such as miombo woodlands in Tanzania. This highlights the need for a robust approach, like a spatially balanced sampling (Benedetti et al., \u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e2017\u003c/span\u003e; Deville \u0026amp; Tille, \u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e2004\u003c/span\u003e; Grafstr\u0026ouml;m \u0026amp; Schelin, \u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e2014\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eA spatially balanced sample ensures that the distribution of auxiliary variables matches the sampling frame\u0026rsquo;s, allowing the random sample to accurately represent the forest conditions concerning the chosen auxiliary variables. This balanced selection of tracts enhances estimation precision, meaning that the estimates are closer to the true values than traditional sampling methods. The design is also flexible, allowing scalability and the inclusion of uncommon and common phenomena within the same framework through different sampling densities, multistage sampling, and the exclusion of tracts (Grafstr\u0026ouml;m \u0026amp; Matei, \u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e2018\u003c/span\u003e). The potential of a balanced sampling design has been increasingly recognized for large-scale environmental monitoring and forest cover change assessment, as demonstrated in studies by Kermorvant et al. (\u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e2019\u003c/span\u003e) and Pagliarella et al. (\u003cspan citationid=\"CR41\" class=\"CitationRef\"\u003e2018\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eThis study employed a multi-level sampling approach to estimate miombo woodland canopy cover in Tanzania. Specifically, we demonstrate the use of a two-stage sampling design combined with spatially balance sampling to provide a precise estimate of canopy cover within the miombo woodland of Western Tanzania.\u003c/p\u003e"},{"header":"Materials and methods","content":"\u003cp\u003eStudy area\u003c/p\u003e \u003cp\u003eWe conducted the study in the Mlele and Tanganyika districts in the Katavi Region, as well as in the Sikonge and Tabora districts in the Tabora Region, all located in Western Tanzania (Table\u0026nbsp;\u003cspan refid=\"Tab6\" class=\"InternalRef\"\u003e1\u003c/span\u003e, Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e1\u003c/span\u003e). The miombo woodlands in the districts experience unimodal rainfall pattern starting in November and ending in April of the following year. The dominant soil type is Cambisols. The climate and soil support both dry and wet miombo woodland vegetation types. The miombo canopy cover in the study area was estimated at 40,163 km\u003csup\u003e2\u003c/sup\u003e in 2009 during the National Forestry Resources Monitoring and Assessment (NAFORMA) (URT, \u003cspan citationid=\"CR52\" class=\"CitationRef\"\u003e2015\u003c/span\u003e). Subsequent estimates indicate variations in tree cover for the study area, with values of 49,195 km\u003csup\u003e2\u003c/sup\u003e in 2018 from the Verhegghen et al. (\u003cspan citationid=\"CR53\" class=\"CitationRef\"\u003e2022\u003c/span\u003e) tree cover/non-tree cover dataset, 30,035 km\u003csup\u003e2\u003c/sup\u003e in 2020 from the ESA World cover dataset (Zanaga et al., \u003cspan citationid=\"CR55\" class=\"CitationRef\"\u003e2021\u003c/span\u003e), and 41,710 km\u003csup\u003e2\u003c/sup\u003e in 2021 from the ESRI land cover dataset (Karra et al., \u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e2021\u003c/span\u003e).\u003c/p\u003e \u003cp\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\u003eDescription of districts comprising the study area\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"8\"\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=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c7\" colnum=\"7\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c8\" colnum=\"8\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003eDistrict\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003eLand area (km\u003csup\u003e2\u003c/sup\u003e)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colspan=\"2\" nameend=\"c4\" namest=\"c3\"\u003e \u003cp\u003eLocation range\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003eElevation range (masl)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colspan=\"2\" nameend=\"c7\" namest=\"c6\"\u003e \u003cp\u003eClimate ranges\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c8\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003eMiombo type\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eLatitude\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eLongitude\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c6\"\u003e \u003cp\u003eRainfall (mm)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c7\"\u003e \u003cp\u003eTemperature (℃)\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTabora\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e1460\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e4\u0026deg; 40' 12''S \u0026minus;\u0026thinsp;5\u0026deg; 12' 47''S\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e32\u0026deg; 38' 20''E \u0026minus;\u0026thinsp;32\u0026deg; 59' 38''E\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e1123\u0026ndash;1394\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e916\u0026ndash;1008\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e23\u0026ndash;24\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003eDry\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSikonge\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e26260\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e5\u0026deg; 11' 42''S \u0026minus;\u0026thinsp;6\u0026deg; 58' 34''S\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e32\u0026deg; 01' 44''E \u0026minus;\u0026thinsp;34\u0026deg; 06' 36''E\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e1059\u0026ndash;1632\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e648\u0026ndash;1010\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e20\u0026ndash;24\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003eDry\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMlele\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e15558\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e5\u0026deg; 42' 47''S \u0026minus;\u0026thinsp;7\u0026deg; 16' 08''S\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e31\u0026deg; 25' 30''E \u0026minus;\u0026thinsp;32\u0026deg; 45' 25''E\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e832\u0026ndash;1670\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e858\u0026ndash;1054\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e20\u0026ndash;25\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003eDry and Wet\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTanganyika\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e16932\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e5\u0026deg; 07' 59''S \u0026minus;\u0026thinsp;6\u0026deg; 56' 13''S\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e29\u0026deg; 54' 11''E \u0026minus;\u0026thinsp;31\u0026deg; 11' 10''E\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e759\u0026ndash;2034\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e942\u0026ndash;1244\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e19\u0026ndash;26\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003eWet\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\u003eSampling design\u003c/p\u003e \u003cp\u003eWe utilized a two-stage sampling design that incorporated spatially balanced sampling to select 1 km x 1 km tracts in the first stage, followed by stratified random sampling that chose 30 m x 30 m plots in the second stage. The processes involved in the selection of tracts and plots are as follows:\u003c/p\u003e \u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003eSelection of 1 km x 1 km tracts in the first stage\u003c/h2\u003e \u003cp\u003eWe generated a sampling frame of 998,925 tracts of 1 km x 1 km that covered the entire country and prepared auxiliary variables (as listed in Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e) that characterized miombo woodland distribution in Tanzania. The preparation of auxiliary variables involved collecting data from the specified sources and harmonizing their coordinate reference systems and spatial resolutions. The coordinate reference system of the tracts was also matched with those of auxiliary data so that they could overlay.\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\u003eAuxiliary variables used in the balance sampling.\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\u003eAuxiliary variable name\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eJustification/Explanation\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eSource\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eData description\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eWoodland cover\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eIt depicted the distribution of miombo woodland in Tanzania for the year 2019. The layer was generated by subtracting deforested areas between 2013 and 2019, as indicated in Hansen's loss year 2019, from the National Land Cover 2013. The processing was conducted using ArcMap software.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eThe National land cover of 2013 and the Hansen dataset for 2019 (Loss year) (Hansen et al., \u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e2023\u003c/span\u003e)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eMiombo woodland distribution in Tanzania at 30m resolution\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNDVI pattern\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eIt measures vegetation health and can be used to distinguish between different vegetation types.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eMODIS Q1\u003c/p\u003e \u003cp\u003eVegetation indices (NDVI and EVI 16-Day L3 Global 250m)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eVegetation indices for August 2021\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eElevation\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eIt is one of the sources of heterogeneity in miombo woodlands (Mwakalukwa et al., \u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e2014\u003c/span\u003e).\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eJAXA DEM (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.eorc.jaxa.jp/ALOS/en/aw3d30/data/index.htm\u003c/span\u003e\u003cspan address=\"https://www.eorc.jaxa.jp/ALOS/en/aw3d30/data/index.htm\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eDigital Elevation Model 30 m resolution\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eRainfall\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e● It is used to characterize miombo woodlands into dry and wet types (Jinga \u0026amp; Palagi, \u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e2020\u003c/span\u003e; Lilleso et al., \u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e2014\u003c/span\u003e).\u003c/p\u003e \u003cp\u003e● It is an environmental factor affecting the distribution of miombo tree species (Chidumayo, \u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e2017\u003c/span\u003e; Jinga \u0026amp; Palagi, \u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e2020\u003c/span\u003e).\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eWorldClim precipitation data at 2.5 m (21 km\u003csup\u003e2\u003c/sup\u003e) resolution 1970\u0026ndash;2000.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eMonthly precipitation (mm). The data were summarized to obtain \u0026ndash;total annual rainfall.\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSoil Moisture Index\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eIt is related to rainfall and is an environmental factor affecting the distribution of miombo tree species.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eSMOS L2 Science data 2500 m was obtained from the ESA site. (\u003c/p\u003e \u003cp\u003e\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://smos-diss.eo.esa.int/smos/faq.html#Available%20data\u003c/span\u003e\u003cspan address=\"https://smos-diss.eo.esa.int/smos/faq.html#Available%20data\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eSoil Moisture (m3m-3) for August 2021\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSoil Organic Carbon Stock\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eIt is a soil nutrient that determines miombo species composition (Mwakalukwa et al., \u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e2014\u003c/span\u003e).\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eFAO GLOSIS at 500 m resolution - Global Soil Organic Carbon Map (v1.5.0). ( \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttp://54.229.242.119/GSOCmap\u003c/span\u003e\u003cspan address=\"http://54.229.242.119/GSOCmap\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eSoil Organic Carbon in tonnes/ha\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTotal nitrogen\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eIt is a soil nutrient that determines miombo species composition (Mwakalukwa et al., \u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e2014\u003c/span\u003e).\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eISRIC World Soil Information at 250 m resolution. ( \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://data.isric.org/geonetwork/srv/eng/catalog.search#/metadata/c5cb0f71-a627-40b9-8acb-5be42e4d5914\u003c/span\u003e\u003cspan address=\"https://data.isric.org/geonetwork/srv/eng/catalog.search#/metadata/c5cb0f71-a627-40b9-8acb-5be42e4d5914\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eTotal Nitrogen (N) content in mg/kg (ppm) spatially predicted for 0\u0026ndash;30 cm depth.\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTotal phosphorus\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eIt is one of the soil nutrients that determine miombo species composition (Shirima, \u003cspan citationid=\"CR47\" class=\"CitationRef\"\u003e2015\u003c/span\u003e).\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eISRIC World Soil Information at 250 m resolution. (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://data.isric.org/geonetwork/srv/eng/catalog.search#/metadata/f72f5698-4a3d-4af2-ae12-a4b1cf151cec\u003c/span\u003e\u003cspan address=\"https://data.isric.org/geonetwork/srv/eng/catalog.search#/metadata/f72f5698-4a3d-4af2-ae12-a4b1cf151cec\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eTotal Phosphorus (P) content in mg/kg (ppm) for 0\u0026ndash;30 cm depth.\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\u003eWe extracted values from each auxiliary layer for every tract. These extracted values were then used as inputs for the Balanced Sampling tool in R software (Grafstr\u0026ouml;m \u0026amp; Lisic, \u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e2019\u003c/span\u003e), which sampled 1000 tracts across Tanzania, ensuring equal inclusion probabilities. Of the 998,925 tracts that covered Tanzania, 60,210 tracts covered the study area. Further, of the 1000 tracts sampled by the Balanced Sampling tool, 68 were located within our study area (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec4\" class=\"Section2\"\u003e \u003ch2\u003eSelection of 30 m x 30 m plots in the second stage\u003c/h2\u003e \u003cp\u003eEach of the selected 68 tracts comprised an average of 1025 plots of 30 m x 30 m, categorized based on the percentage of woodland cover, as shown in Table\u0026nbsp;\u003cspan refid=\"Tab3\" class=\"InternalRef\"\u003e3\u003c/span\u003e. The miombo woodland cover was determined by classifying Sentinel-2 images from 2020/2021 using the random forest method (Appendix 1). The definitions for these woodland cover classes were adopted from the United Nations Environment Program (UNEP), which classifies 'closed canopy' as having a canopy cover of more than 40% and 'open canopy' as having a canopy cover of 10\u0026ndash;40%. The Tropical Ecosystem Environment Observations by Satellite (TREES) project also defines a 'dense canopy' with more than 70% canopy cover. (Tchatchou et al., \u003cspan citationid=\"CR51\" class=\"CitationRef\"\u003e2015\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eWe applied stratified random sampling to select a maximum of 40 plots from the categorized 1025 plots within each of the selected 68 tracts. Table\u0026nbsp;\u003cspan refid=\"Tab3\" class=\"InternalRef\"\u003e3\u003c/span\u003e details the number of plots in the sampling frame and the number of plots selected for each canopy cover category in the sampling frame.\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\u003eNumber of plots and number of sampled plots per cover categories in selected tracts\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\u003eCover categories\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNumber of plots\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eNumber of sampled plots\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eClosed miombo woodland (canopy cover\u0026thinsp;\u0026gt;\u0026thinsp;70% )\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e32442\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e1338\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eOpen miombo woodland (40% \u0026le; canopy cover\u0026thinsp;\u0026le;\u0026thinsp;70%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e6683\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e201\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eVery open miombo woodland (10% \u0026le; canopy cover\u0026thinsp;\u0026lt;\u0026thinsp;40%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e7901\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e228\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNon-miombo woodlands (canopy cover\u0026thinsp;\u0026lt;\u0026thinsp;10%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e22690\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e923\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eTotal\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e\u003cb\u003e69716\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e\u003cb\u003e2690\u003c/b\u003e\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\u003ePreparation of reference datasets\u003c/p\u003e \u003cp\u003eIn this study, we generated classified maps of woodland and non-woodland for each tract by analyzing Planetscope imagery. These maps served as reference datasets for collecting canopy cover data per plot. The decision to utilize remote sensing data, specifically Planetscope imagery, was prompted by its cost-effectiveness compared to ground data collection. Additionally, remote sensing offers a standardized approach to measuring canopy cover, minimizing variations from different observers and collection times (seasons) associated with ground-based methods. In contrast to wall-to-wall mapping, image classification at the tract level is more accurate due to more intensive effort applied to smaller tracts (cf. Stehman et al. (\u003cspan citationid=\"CR49\" class=\"CitationRef\"\u003e2003\u003c/span\u003e).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec5\" class=\"Section2\"\u003e \u003ch2\u003ePlanetScope basemaps acquisition\u003c/h2\u003e \u003cp\u003ePlanetScope basemaps, Normalized Analytic, were acquired from the National Imaging and Mapping Agency (NICI) (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.planet.com/basemaps/#/mode/selection/mosaic\u003c/span\u003e\u003cspan address=\"https://www.planet.com/basemaps/#/mode/selection/mosaic\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003cspan type=\"Underline\" class=\"Underline\" name=\"Emphasis\"\u003e).\u003c/span\u003e PlanetScope was selected for its high spatial resolution (3\u0026ndash;5 m), making its images suitable for visual interpretation (Asrat et al., \u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e2018\u003c/span\u003e), and its daily availability, enhancing the probability of obtaining monthly cloud-free images.\u003c/p\u003e \u003cp\u003eBefore selecting the monthly images, we analyzed the bands for each month between January 2022 and December 2022. The analysis showed November and December 2022 as the most suitable months for mapping miombo vegetation in the study area. This choice aligned with the timing of miombo species leaf bud development, which began in August before the October rains. In November, miombo trees in the lowlands exhibited full leaf canopies, and by December, the majority of miombo tree species in both lowlands and highlands displayed full leaf canopies. Additionally, in November and December, the presence of crops on farms and the occurrence of burned areas had minimal effects on spectral signals.\u003c/p\u003e \u003cp\u003eThus, we utilized the imagery from November and December 2022. Additionally, January 2023 images were included to increase the likelihood of obtaining cloud-free imagery.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec6\" class=\"Section2\"\u003e \u003ch2\u003ePlanetScope imageries processing\u003c/h2\u003e \u003cp\u003eMapping miombo woodland from the collected planetScope images was done through Normalized Difference Vegetation Index (NDVI) values, followed by a visual verification step. NDVI layers were generated using the formula (1), and the miombo woodland class was separated by NDVI thresholds that ranged between 0.5 and 0.65. The overall classification accuracy of PlanetScope imagery within tracts was 0.94, accompanied by a Kappa statistic of 0.61 (Appendix 2).\u003cdiv id=\"Equa\" class=\"Equation\"\u003e\u003cdiv format=\"TEX\" class=\"mathdisplay\" id=\"FileID_Equa\" name=\"EquationSource\"\u003e\n$$NDVI= \\frac{NIR-Red}{NIR+Red} \\left(1\\right)$$\u003c/div\u003e\u003c/div\u003e\u003c/p\u003e \u003cp\u003eWhere\u003c/p\u003e \u003cp\u003eNIR\u0026thinsp;=\u0026thinsp;Near Infrared band and\u003c/p\u003e \u003cp\u003eRed\u0026thinsp;=\u0026thinsp;Red band\u003c/p\u003e \u003cp\u003eData collection\u003c/p\u003e \u003cp\u003eData were collected in sampled 30 m x 30 m plots within each tract. The miombo woodland class, prepared for each tract as described above, was overlaid with the sampled plots. Subsequently, the canopy cover, representing the percentage of the miombo woodland in the plot, was determined in each plot. This implies that canopy cover was assessed as viewed from above by PlanetScope (cf. Westfall \u0026amp; Morin, \u003cspan citationid=\"CR54\" class=\"CitationRef\"\u003e2013\u003c/span\u003e). This process is also known as diffuse canopy cover measurements, wherein overlapping crowns from neighboring trees are combined.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec7\" class=\"Section2\"\u003e \u003ch2\u003eData analysis\u003c/h2\u003e \u003cdiv id=\"Sec8\" class=\"Section3\"\u003e \u003ch2\u003eCanopy Cover Estimation\u003c/h2\u003e \u003cp\u003eThe Horvitz\u0026ndash;Thompson (H.T.) estimator was chosen for its compatibility with the Spatial Balanced Design (Grafstr\u0026ouml;m \u0026amp; Till\u0026eacute;, \u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e2013\u003c/span\u003e). Using equations (1)\u0026ndash;(3) in Appendix (3), the estimator determined the total area of miombo woodland with closed, open, and very open canopy covers based on the number of plots in each canopy cover category, as identified from the classified PlanetScope images. Subsequently, equations (4)\u0026ndash;(8) of the H.T. estimator (in Appendix 3) were used to estimate the canopy cover within the miombo woodland for closed, open, and very open canopy covers, utilizing the collected data on canopy cover. The standard error was computed as the square root of the variance estimated from the equations, and the relative standard error (RSE) was calculated as the ratio of the standard error to the estimated mean or total.\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e"},{"header":"Results and Discussion","content":"\u003ch3\u003eResults\u003c/h3\u003e\n\u003cdiv id=\"Sec11\" class=\"Section2\"\u003e \u003ch2\u003eCanopy cover estimates\u003c/h2\u003e \u003cp\u003eThe findings reveal that miombo woodland coverage within our study area was 37,359\u0026thinsp;\u0026plusmn;\u0026thinsp;4,618 km\u0026sup2;, representing approximately 54\u0026ndash;70% of the study area (Table\u0026nbsp;\u003cspan refid=\"Tab4\" class=\"InternalRef\"\u003e4\u003c/span\u003e). Among different miombo woodland classes, the closed canopy cover type contributed the largest portion, covering an area of 29,546\u0026thinsp;\u0026plusmn;\u0026thinsp;4,382 km\u0026sup2;, constituting about 42\u0026ndash;56% of the study area. The very open canopy cover type followed closely, with an area of 4,094\u0026thinsp;\u0026plusmn;\u0026thinsp;1,127 km\u0026sup2;, making up approximately 5\u0026ndash;9% of the study area. The smallest contribution was from the open canopy cover type, with an area of 3,719\u0026thinsp;\u0026plusmn;\u0026thinsp;849 km\u0026sup2;, accounting for about 5\u0026ndash;8% of the study area.\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\u003eEstimates of miombo woodland area (in square kilometers) for 2022/2023\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=\"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 \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCanopy cover class\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eArea (km\u003csup\u003e2\u003c/sup\u003e)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eME (\u0026plusmn;\u0026thinsp;95% confidence interval)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eRSE (%)\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eThe miombo woodland area with closed canopy cover (cover% \u0026gt; 70%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e29546\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e4382\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e7.57\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eThe miombo woodland area with open canopy cover (40% \u0026le; cover% \u0026le; 70%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e3719\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e849\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e11.64\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eThe miombo woodland area with very open canopy cover (10% \u0026le; cover% \u0026lt; 40%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e4094\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e1127\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e14.04\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eThe miombo woodland area (\u0026ge;\u0026thinsp;10%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e37359\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e4618\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e6.31\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\u003eFurthermore, the findings indicate that the canopy cover in the miombo woodland area was 55% \u0026plusmn; 5%. Consequently, the miombo woodland in the study area predominantly fell within the open canopy cover class. Regarding miombo woodland classes, the canopy cover within the areas classified as closed, open, and very open cover classes are presented in Table\u0026nbsp;\u003cspan refid=\"Tab5\" class=\"InternalRef\"\u003e5\u003c/span\u003e below.\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\u003eEstimates of canopy cover in the miombo woodland area\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=\"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 \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCanopy cover class\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eCanopy cover (%)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eME (\u0026plusmn;\u0026thinsp;95% confidence interval)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eRSE (%)\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eThe canopy cover in the miombo woodland area with closed canopy cover (cover% \u0026gt; 70%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e84\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e7\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e4.50\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eThe canopy cover in the miombo woodland area with open canopy cover (40% \u0026le; cover% \u0026le; 70%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e45\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e6\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e6.34\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eThe canopy cover in the miombo woodland area with very open canopy cover (10% \u0026le; cover% \u0026lt; 40%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e19\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e6.91\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eThe canopy cover in the miombo woodland area (\u0026ge;\u0026thinsp;10%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e55\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e4.63\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\n\u003ch3\u003eDiscussion\u003c/h3\u003e\n\u003cp\u003eIn this paper, we demonstrate the design of a monitoring scheme for miombo woodland in Tanzania, producing accurate and precise estimates of its areal extent and canopy cover in the study area. Notably, all estimates displayed Relative Standard Errors (RSEs) well below 25%, indicating their reliability, as Sileshi (\u003cspan citationid=\"CR48\" class=\"CitationRef\"\u003e2015\u003c/span\u003e) suggested.\u003c/p\u003e \u003cp\u003eThe results reveal that the miombo woodland in the study area covers 37,359\u0026thinsp;\u0026plusmn;\u0026thinsp;4,618 km\u003csup\u003e2\u003c/sup\u003e, with an average canopy cover of 55% \u0026plusmn; 5%. The closed miombo woodland, with an average canopy cover of 84% \u0026plusmn; 7%, comprised 79% of the miombo woodland area. Open miombo woodland, with an average canopy cover of 45% \u0026plusmn; 6%, and very open miombo woodland, with an average canopy cover of 19% \u0026plusmn; 3%, comprised 10% and 11% of the miombo woodland area, respectively. This could be attributed to the fact that 74% of the miombo woodland in the study area is within protected areas and influenced by the presence of wet miombo woodlands. A closed canopy characterizes wet miombo woodland due to a positive correlation between rainfall and miombo growth rates (Frost, \u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e1996\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eHowever, the results regarding canopy cover do not align with the estimates provided by the Moderate Resolution Imaging Spectroradiometer vegetation continuous fields (MODIS VCF) tree cover product for the year 2022 (cf. DiMiceli et al., \u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e2017\u003c/span\u003e). According to the MODIS VCF data, very open miombo woodland, open miombo woodland, and closed miombo woodland covers an area of 40,680 km\u003csup\u003e2\u003c/sup\u003e, 2,557 km\u003csup\u003e2\u003c/sup\u003e, and 6 km\u003csup\u003e2\u003c/sup\u003e, respectively, totaling 43,243 km\u003csup\u003e2\u003c/sup\u003e of the miombo woodland area. This suggests a dominance of very open miombo woodlands, according to MODIS VCF. Contrastingly, the findings of Adzhar et al. (\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e2021\u003c/span\u003e) indicate that MODIS VCF consistently underestimates woody cover in tropical savannas. The observed discrepancy raises questions about the accuracy of MODIS VCF estimates, particularly in characterizing miombo woodland cover in the studied area.\u003c/p\u003e \u003cp\u003eThe area of miombo woodland found in this study align with the estimates from the ESRI land cover dataset that revealed the trees covered 41,710 km\u003csup\u003e2\u003c/sup\u003e in 2021 (Karra et al., \u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e2021\u003c/span\u003e). Other recent land cover assessment estimates revealed the trees covered 49,195 km\u003csup\u003e2\u003c/sup\u003e in 2018 (Verhegghen et al., \u003cspan citationid=\"CR53\" class=\"CitationRef\"\u003e2022\u003c/span\u003e) and 30,035 km\u003csup\u003e2\u003c/sup\u003e in 2020 (Zanaga et al., \u003cspan citationid=\"CR55\" class=\"CitationRef\"\u003e2021\u003c/span\u003e). However, the estimates of Karra et al. (\u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e2021\u003c/span\u003e), Verhegghen et al. (\u003cspan citationid=\"CR53\" class=\"CitationRef\"\u003e2022\u003c/span\u003e), and Zanaga et al. (\u003cspan citationid=\"CR55\" class=\"CitationRef\"\u003e2021\u003c/span\u003e) from wall-to-wall mapping lack standard errors, contributing to difficulty in assessing the reliability and precision of their estimates for better comparison with our study estimates.\u003c/p\u003e \u003cp\u003eWall-to-wall mapping, which is suitable for providing spatially explicit information, contains errors arising from data, classification algorithms, and analysts that affect the accuracy of land cover estimations. To address these uncertainties, evaluations of wall-to-wall mapping often involve comparing the produced land cover with ground truth data using a confusion matrix. In order to offer an unbiased assessment of wall-to-wall mapping, a design-based statistical sampling is employed to estimate land cover with confidence levels (Corona, \u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e2010\u003c/span\u003e; Corona et al., \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e2015\u003c/span\u003e). This approach helps mitigate errors and provides a more robust estimation framework (cf. Sannier et al. \u003cspan citationid=\"CR46\" class=\"CitationRef\"\u003e2014\u003c/span\u003e). Additionally, a sample-based approach to land cover estimation proves to be cost-effective and time-efficient, as efforts are concentrated in a smaller representative area.\u003c/p\u003e \u003cp\u003eDue to the estimates of miombo woodland obtained in this study and the advantages of statistical sampling, our results support the idea that wall-to-wall mapping should be integrated with sample-based assessment (cf. Maniatis et al. \u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e2021\u003c/span\u003e and Gallaun et al. \u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e2015\u003c/span\u003e). This integration will help reduce uncertainties resulting from the challenges associated with dryland canopy cover monitoring, as Bodart et al. (\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e2013\u003c/span\u003e) observed.\u003c/p\u003e \u003cp\u003eFurthermore, our results indicate an improvement in the reliability of estimates with increasing canopy cover, as indicated by the decreasing Relative Standard Error (RSE) from very open to open and closed canopy cover classes. This improvement is linked to sample size, where a closed woodland class was the most common, and the homogeneity of measured canopy cover within the class. Despite a smaller sample size within the open woodland class compared to the very open woodland class, the RSE within the open woodland class was lower, reflecting reduced variability in the measured canopy cover within this class. Corona et al. (\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e2015\u003c/span\u003e) similarly found that the greatest sampling effort resulted in a low Relative Standard Error (RSE), especially with many primary sampling units (n) at the first stage and few secondary sampling units (m) at the second stage, for a specific sample size (m x n) in the two-stage sampling strategy. Therefore, increasing the sample size, primarily through increasing the number of tracts in the first stage, will likely enhance the accuracy of estimates derived from the sampling strategy.\u003c/p\u003e \u003cp\u003eThe inventory design used in this study is based on the inventory design used in the national monitoring program NILS \u0026ndash; National Inventories of Landscapes in Sweden (Allard et al. \u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e2023\u003c/span\u003e, Brown et al. \u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e2023\u003c/span\u003e, Ranlund et al. \u003cspan citationid=\"CR45\" class=\"CitationRef\"\u003e2023\u003c/span\u003e; \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.slu.se/en/Collaborative-Centres-and-Projects/nils/\u003c/span\u003e\u003cspan address=\"https://www.slu.se/en/Collaborative-Centres-and-Projects/nils/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e). The methodology involves among other things a multi-stage sampling approach. In the first stage, we utilize spatially balanced sampling to select tracts, and in the second stage, we utilize stratified random sampling to select plots within the tracts. This two-stage sampling approach is effective since only 0.1% of the tracts and 4% of the plots need to be inventoried.\u003c/p\u003e \u003cp\u003eHowever, the tracts and plots that did not need to be inventoried in the planetScope in the second stage were included, contributing to the large sample size and high precision of the estimates. These were tracts and plots lacking the focal phenomena (in this case, miombo woodland). It was possible to include tracks and plots lacking focal phenomena, as the study's second stage involved an ocular inventory of selected plots. However, for costly and time-consuming field inventories, the focus can be on tracts and plots with the potential miombo woodland or with the potential miombo canopy cover class to ensure cost-effectiveness.\u003c/p\u003e \u003cp\u003eFurther, the design allows for different sampling methods in the second stage and between primary sampling units for the selection of secondary sampling units within them. The choice of methods is guided by the characteristics of the population and research objectives, which also might result in varying sample sizes of secondary sampling units between primary sampling units. This inherent flexibility positions the design as a potential model for an effective national monitoring program, seamlessly integrating remote sensing and field inventories.\u003c/p\u003e"},{"header":"Conclusions","content":"\u003cp\u003eLong-term monitoring of miombo woodlands is crucial to enhance their management, given the growing pressure from unsustainable harvesting and the impacts of climate and land use changes. This study established the reliability and precision of a monitoring design employing a two-stage sampling strategy that utilizes spatially balanced sampling in the first stage. Results from this approach revealed that miombo woodlands in Tabora, Sikonge, Mlele, and Tanganyika districts collectively cover 37,359\u0026thinsp;\u0026plusmn;\u0026thinsp;4,618 km\u0026sup2;, with an average canopy cover of 55% \u0026plusmn; 5%. Moreover, closed miombo woodland, with an average canopy cover of 84% \u0026plusmn; 7%, dominates the landscape, covering an area of 29,546\u0026thinsp;\u0026plusmn;\u0026thinsp;4,382 km\u0026sup2;. These results had relative standard errors consistently below 25%. These findings on miombo woodlands' canopy cover provide a baseline for monitoring miombo types and studying various ecological processes. The design exhibits effectiveness by enabling focused sampling and demonstrates flexibility by allowing the integration of varying sampling methods in the second stage. Furthermore, the proposed sampling strategy could potentially serve as a valuable tool for validating wall-to-wall mapping efforts and as an effective model for a national canopy monitoring program.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003eInterests: The authors have no competing interests to declare that are relevant to the content of this article.\u003c/p\u003e\u003cp\u003eAll authors have read, understood, and have complied as applicable with the statement on \"Ethical responsibilities of Authors' as found in the Instructions for Authors.\u003c/p\u003e\u003ch2\u003eFunding:\u003c/h2\u003e \u003cp\u003eThis work was funded by Sida under Grant Agreement No 13394.\u003c/p\u003e\u003ch2\u003eAuthor Contribution\u003c/h2\u003e\u003cp\u003eSiwa Ernest Nkya, Deo Dominick Shirima, Henrik Hedenas, and August Temu contributed to the design and implementation of the research. Prior to writing the manuscript, Siwa Ernest Nkya and Henrik Hedenas analyzed the data. In the manuscript, Siwa Ernest Nkya and Henrik Hedenas wrote the introduction, results, and discussion sections; Siwa Ernest Nkya and Deo Dominick Shirima wrote the methodology section, and Siwa Ernest Nkya and August Temu wrote the abstract and conclusion sections. All authors revised the manuscript and approved its submission for publication to Environment Monitoring and Assessment.\u003c/p\u003e\u003ch2\u003eData availability:\u003c/h2\u003e \u003cp\u003eThe data supporting the findings of this study are available within the paper and its supplementary information files.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eAbdallah, J., \u0026amp; Monela, G.G. (2007). Overview of Miombo woodlands in Tanzania. \u003cem\u003eProceedings of the First MITMIOMBO Project Work- Shop\u003c/em\u003e, \u003cem\u003e50\u003c/em\u003e(February), 9\u0026ndash;23. \u003c/li\u003e\n\u003cli\u003ehttps://jukuri.luke.fi/bitstream/handle/10024/535965/mwp050.pdf?isAllowed=y\u0026amp;sequence=1\u003c/li\u003e\n\u003cli\u003eAdzhar, R., Kelley, D., Dong, N., Torello Raventos, M., Veenendaal, E., Feldpausch, T., Philips, O., Lewis, S., Sonk\u0026eacute;, B., Taedoumg, H., Schwantes Marimon, B., Domingues, T., Arroyo, L., Djagbletey, G., Saiz, G., \u0026amp; Gerard, F. (2021). Assessing MODIS Vegetation Continuous Fields tree cover product (collection 6): performance and applicability in tropical forests and savannas. Biogeosciences Discussions, February, 1\u0026ndash;20. https://doi.org/10.5194/bg-2020-460\u003c/li\u003e\n\u003cli\u003eAllard, A., Brown, A., Hurford, C., Isendahl, C., Hilpold, A., Pappeiner, U., Strolb J. \u0026amp; Heden\u0026aring;s, H. 2023. Chapter 5. Data collected in situ, unique details or integrated components of monitoring schemes. In. Allard, A., Keskitalo, C.H. \u0026amp; Brown A. (reds). Monitoring Biodiversity Combining Environmental and Social Data. Routledge. \u003c/li\u003e\n\u003cli\u003eAsrat, Z., Taddese, H., \u0026Oslash;rka, H. O., Gobakken, T., \u0026amp; N\u0026aelig;sset, E. (2018). Estimation of Forest Area and Canopy Cover Based on Visual Interpretation of Satellite Images in Ethiopia. \u003cem\u003eLand\u003c/em\u003e, \u003cem\u003e7\u003c/em\u003e(92), 1\u0026ndash;17. https://doi.org/10.3390/land7030092\u003c/li\u003e\n\u003cli\u003eBenedetti, R., Piersimoni, F., \u0026amp; Postiglione, P. (2017). Alternative and complementary approaches to spatially balanced samples. \u003cem\u003eMetron\u003c/em\u003e, \u003cem\u003e75\u003c/em\u003e(3), 249\u0026ndash;264. https://doi.org/10.1007/s40300-017-0123-1\u003c/li\u003e\n\u003cli\u003eBodart, C., Brink, A. B., Donnay, F., Lupi, A., Mayaux, P., \u0026amp; Achard, F. (2013). Continental estimates of forest cover and forest cover changes in the dry ecosystems of Africa between 1990 and 2000. \u003cem\u003eJournal of Biogeography\u003c/em\u003e, \u003cem\u003e40\u003c/em\u003e(6), 1036\u0026ndash;1047. https://doi.org/10.1111/jbi.12084\u003c/li\u003e\n\u003cli\u003eBrown, A., Heden\u0026aring;s, H., Holm, E., Lind, T., Richards, A.E., Prober, S.M. \u0026amp; Schmidt B. 2023. Chapter 15. Designing and Adapting Biodiversity Monitoring Schemes. In. Allard, A., Keskitalo, C.H. \u0026amp; Brown A. (reds). Monitoring Biodiversity Combining Environmental and Social Data. Routledge.\u003c/li\u003e\n\u003cli\u003eChidumayo, E. N. (2001). Climate and Phenology of Savanna Vegetation in Southern Africa. Journal of Vegetation Science, 12(3), 347. https://doi.org/10.2307/3236848\u003c/li\u003e\n\u003cli\u003eChidumayo, E. N. (2017). \u003cem\u003eBiotic interactions, climate and disturbance underlie the distribution of two Julbernardia tree species in miombo woodlands of Africa\u003c/em\u003e. 1\u0026ndash;11. https://doi.org/10.1017/S0266467416000584\u003c/li\u003e\n\u003cli\u003eChidumayo, E. N. (2019). Management implications of tree growth patterns in miombo woodlands of Zambia. Forest Ecology and Management, 436(October 2018), 105\u0026ndash;116. https://doi.org/10.1016/j.foreco.2019.01.018\u003c/li\u003e\n\u003cli\u003eCorbera, E., Estrada, M., \u0026amp; Brown, K. (2010). Reducing greenhouse gas emissions from deforestation and forest degradation in developing countries: Revisiting the assumptions. \u003cem\u003eClimatic Change\u003c/em\u003e, \u003cem\u003e100\u003c/em\u003e(3), 355\u0026ndash;388. https://doi.org/10.1007/s10584-009-9773-1\u003c/li\u003e\n\u003cli\u003eCorona, P. (2010). Integration of forest mapping and inventory to support forest management. \u003cem\u003eIForest - Biogeosciences and Forestry\u003c/em\u003e, 59\u0026ndash;64. https://doi.org/10.3832/ifor0531-003\u003c/li\u003e\n\u003cli\u003eCorona, P., Fattorini, L., \u0026amp; Pagliarella, M. C. (2015). Sampling strategies for estimating forest cover from remote sensing-based two-stage inventories. \u003cem\u003eForest Ecosystems\u003c/em\u003e, \u003cem\u003e2\u003c/em\u003e(1). https://doi.org/10.1186/s40663-015-0042-7\u003c/li\u003e\n\u003cli\u003eDeville, B. J., \u0026amp; Tille, Y. (2004). Efficient\u003cem\u003e balanced sampling : The cube method\u003c/em\u003e. 893\u0026ndash;912.\u003c/li\u003e\n\u003cli\u003eDewees, P. ., Campbell, B. ., Katerere, Y., Sitoe, A., Cunningham, A. ., Angelsen, A., \u0026amp; Wunder, S. (2011). \u003cem\u003eManaging the Miombo Woodlands of Southern Africa: Policies, incentives, and options for the rural poor\u003c/em\u003e. Program on Forests (PROFOR).\u003c/li\u003e\n\u003cli\u003eDiMiceli, C. ., Carroll, M. ., Sohlberg, R. ., Huang, C., Hansen, M. C., \u0026amp; Townshend, J. R. G. (2017). Annual global automated MODIS vegetation continuous fields (MOD44B) at 250 m spatial resolution for data years beginning day 65, 2000-2014, collection 5 percent tree cover, version 6 (p. 12). University of Maryland, College Park, MD. https://lpdaac.usgs.gov/documents/1494/MOD44B_User_Guide_V61.pdf\u003c/li\u003e\n\u003cli\u003eFAO. (2022). FRA 2020 Remote Sensing Survey. In \u003cem\u003eFAO Forestry Paper No. 186\u003c/em\u003e. https://doi.org/10.4060/cb9970en\u003c/li\u003e\n\u003cli\u003eFrost, P. (1996). The ecology of Miombo woodlands. In B. Campbell (Ed.), \u003cem\u003eThe Miombo in Transition: Woodlands and Welfare in Africa\u003c/em\u003e (Issue June, pp. 11\u0026ndash;57). Center for International Forestry Research (CIFOR). http://books.google.com/books?hl=nl\u0026amp;lr=\u0026amp;id=rpildJJVdU4C\u0026amp;pgis=1\u003c/li\u003e\n\u003cli\u003eGallaun, H., Steinegger, M., Wack, R., Schardt, M., Kornberger, B., \u0026amp; Schmitt, U. (2015). Remote sensing based two-stage sampling for accuracy assessment and area estimation of land cover changes. \u003cem\u003eRemote Sensing\u003c/em\u003e, \u003cem\u003e7\u003c/em\u003e(9), 11992\u0026ndash;12008. https://doi.org/10.3390/rs70911992\u003c/li\u003e\n\u003cli\u003eGrafstr\u0026ouml;m, A., \u0026amp; Lisic, J. (2019). Package \u0026lsquo;BalancedSampling .\u0026rsquo; \u003cem\u003eCRAN Reference Manual\u003c/em\u003e, 25. https://cran.r-project.org/web/packages/BalancedSampling/BalancedSampling.pdf\u003c/li\u003e\n\u003cli\u003eGrafstr\u0026ouml;m, A., \u0026amp; Matei, A. (2018). Spatially Balanced Sampling of Continuous Populations. \u003cem\u003eScandinavian Journal of Statistics\u003c/em\u003e, \u003cem\u003e45\u003c/em\u003e(3), 792\u0026ndash;805. https://doi.org/10.1111/sjos.12322\u003c/li\u003e\n\u003cli\u003eGrafstr\u0026ouml;m, A., \u0026amp; Schelin, L. (2014). How to Select Representative Samples. \u003cem\u003eScandinavian Journal of Statistics\u003c/em\u003e, \u003cem\u003e41\u003c/em\u003e(2), 277\u0026ndash;290. https://doi.org/10.1111/sjos.12016\u003c/li\u003e\n\u003cli\u003eGrafstr\u0026ouml;m, A., \u0026amp; Till\u0026eacute;, Y. (2013). Doubly balanced spatial sampling with spreading and restitution of auxiliary totals. \u003cem\u003eEnvironmetrics\u003c/em\u003e, \u003cem\u003e24\u003c/em\u003e(2), 120\u0026ndash;131. https://doi.org/10.1002/env.2194\u003c/li\u003e\n\u003cli\u003eGuerrieri, R., Templer, P., \u0026amp; Magnani, F. (2021). Canopy Exchange and Modification of Nitrogen Fluxes in Forest Ecosystems. Current Forestry Reports, 7, 115\u0026ndash;137. https://doi.org/10.1007/s40725-021-00141-y\u003c/li\u003e\n\u003cli\u003eHansen, M. C., Potapov, P. V., Moore, R., Hancher, M., Turubanova, S. ., Tyukavina, A., Thau, D., Stehman, S. V., Goetz, S. J., Loveland, T. R., Kommareddy, A., Egorov, A., Chini, L., Justice, C. O., \u0026amp; Townshend, J. R. G. (2023). High-Resolution Global Maps of 21st-Century Forest Cover Change. \u003cem\u003eScience 342\u003c/em\u003e, \u003cem\u003e850\u003c/em\u003e(53). http://earthenginepartners.appspot.com/science-2013-global-forest\u003c/li\u003e\n\u003cli\u003eJinga, P., \u0026amp; Palagi, J. (2020). Dry and wet miombo woodlands of south-central Africa respond differently to climate change. \u003cem\u003eEnvironmental Monitoring and Assessment\u003c/em\u003e, \u003cem\u003e192:372\u003c/em\u003e, 1\u0026ndash;16. \u003c/li\u003e\n\u003cli\u003ehttps://doi.org/https://doi.org/10.1007/s10661-020-08342-x\u003c/li\u003e\n\u003cli\u003eJohn, E., Bunting, P., Hardy, A., Roberts, O., Giliba, R., \u0026amp; Silayo, D. S. (2020). Modelling the impact of climate change on Tanzanian forests. \u003cem\u003eDiversity and Distributions\u003c/em\u003e, \u003cem\u003e26\u003c/em\u003e(12), 1663\u0026ndash;1686. \u003c/li\u003e\n\u003cli\u003ehttps://doi.org/10.1111/ddi.13152\u003c/li\u003e\n\u003cli\u003eKarra, K., Kontgis, C., Statman-Weil, Z., Mazzariello, J., Mathis, M., \u0026amp; Brumby, S. (2021). Global land use/land cover with Sentinel-2 and deep learning. \u003cem\u003eIGARSS 2021 IEEE International Geoscience and Remote Sensing Symposium\u003c/em\u003e, \u003cem\u003eapril\u003c/em\u003e.\u003c/li\u003e\n\u003cli\u003eKermorvant, C., D\u0026rsquo;Amico, F., Bru, N., Caill-Milly, N., \u0026amp; Robertson, B. (2019). Spatially balanced sampling designs for environmental surveys. \u003cem\u003eEnvironmental Monitoring and Assessment\u003c/em\u003e, \u003cem\u003e191\u003c/em\u003e(8). https://doi.org/10.1007/s10661-019-7666-y\u003c/li\u003e\n\u003cli\u003eLilleso, J.-P. ., van Breugel, P., Kindt, R., Mbago, F., Moshi, H. ., Ndangalasi, H. ., Uronu, L. O. ., Jamnadass, R., \u0026amp; Graudal, L. (2014). \u003cem\u003ePotential Natural Vegetation of Eastern Africa ( Ethiopia , Kenya , Malawi , Rwanda , Tanzania , Uganda and Zambia ). Volume 10. Atlas and Tree Species Composition for Tanzania\u003c/em\u003e (Vol. 10, Issue 62). Department of Geosciences and Natural Resource Management, University of Copenhagen. https://ign.ku.dk/english/research/forest-nature-biomass/tropical-trees-landscapes/software/vecea/series/filer/Tanzania_Volume_10_web.pdf\u003c/li\u003e\n\u003cli\u003eLindenmayer, D. B., Likens, G. E., Andersen, A., Bowman, D., Bull, C. M., Burns, E., Dickman, C. R., Hoffmann, A. A., Keith, D. A., Liddell, M. J., Lowe, A. J., Metcalfe, D. J., Phinn, S. R., Russell-Smith, J., Thurgate, N., \u0026amp; Wardle, G. M. (2012). Value of long-term ecological studies. \u003cem\u003eAustral Ecology\u003c/em\u003e, \u003cem\u003e37\u003c/em\u003e(7), 745\u0026ndash;757. https://doi.org/10.1111/j.1442-9993.2011.02351.x\u003c/li\u003e\n\u003cli\u003eLiu, Y., Shi, W., Tao, F., Shi, X., \u0026amp; Fu, B. (2023). A global synthesis of multi‐factors affecting water storage capacity in forest canopy, litter and soil layers. Geophysical Research Letters, 50.\u003c/li\u003e\n\u003cli\u003eLupala, Z. J., Lusambo, L. P., \u0026amp; Ngaga, Y. M. (2014). Management, Growth, and Carbon Storage in Miombo Woodlands of Tanzania. \u003cem\u003eInternational Journal of Forestry Research\u003c/em\u003e, \u003cem\u003e2014\u003c/em\u003e, 1\u0026ndash;11. https://doi.org/10.1155/2014/629317\u003c/li\u003e\n\u003cli\u003eLusambo, L. P. (2009). \u003cem\u003eEconomics on the Household energy in miombo woodlands of Eastern and Southern Tanzania\u003c/em\u003e. University of Bangor.\u003c/li\u003e\n\u003cli\u003eManiatis, D., Dionisio, D., Guarnieri, L., Marchi, G., Mollicone, D., D\u0026iacute;az, A. S. P., \u0026amp; Morales, C. (2021). Toward a more representative monitoring of land-use and land-cover dynamics: The use of a sample-based assessment through augmented visual interpretation using open foris collect earth. \u003cem\u003eRemote Sensing\u003c/em\u003e, \u003cem\u003e13\u003c/em\u003e(21). https://doi.org/10.3390/rs13214197\u003c/li\u003e\n\u003cli\u003eMittermeier, R. A., Mittermeier, C. G., Brooks, T. M., Pilgrim, J. D., Konstant, W. R., Fonseca, G. A. B., \u0026amp; Kormos, C. (2003). Wilderness and biodiversity conservation. Proceedings of National Academy of Sciences, 100(18), 10309\u0026ndash;10313. https://doi.org/10.1073/pnas.1732458100\u003c/li\u003e\n\u003cli\u003eMwakalukwa, E. E., Meilby, H., \u0026amp; Treue, T. (2014). Floristic Composition, Structure, and Species Associations of Dry Miombo Woodland in Tanzania. \u003cem\u003eISRN Biodiversity\u003c/em\u003e, \u003cem\u003e2014\u003c/em\u003e, 1\u0026ndash;15. https://doi.org/10.1155/2014/153278\u003c/li\u003e\n\u003cli\u003ePagliarella, M. C., Corona, P., \u0026amp; Fattorini, L. (2018). Spatially-balanced sampling versus unbalanced stratified sampling for assessing forest change: evidences in favour of spatial balance. \u003cem\u003eEnvironmental and Ecological Statistics\u003c/em\u003e, \u003cem\u003e25\u003c/em\u003e(1), 111\u0026ndash;123. https://doi.org/10.1007/s10651-017-0378-y\u003c/li\u003e\n\u003cli\u003eParry, M. ., Canziani, O. ., Palutikof, J. ., \u0026amp; Co-authors. (2007). Technical summary. In M. . Parry, O. . Canziani, J. . Palutikof, P. . van der Linden, \u0026amp; E. Hanson, C.E (Eds.), \u003cem\u003eClimate Change 2007: Impacts, Adaptation and Vulnerability. Contribution of Working Group II to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change\u003c/em\u003e (pp. 23\u0026ndash;78). Cambridge University Press. https://doi.org/10.1016/B978-008044910-4.00250-9\u003c/li\u003e\n\u003cli\u003ePetersen, R., Davis, C., Herold, M., \u0026amp; Sy, V. D. E. (2018). \u003cem\u003eTropical forest monitoring: Exploring the gaps between what is required and what is possible for REDD+ and other initiatives\u003c/em\u003e (Issue June, pp. 1\u0026ndash;12). World Resources Institute. wri.org/ending-tropicaldeforestation\u003c/li\u003e\n\u003cli\u003ePfeifer, M., Gonsamo, A., Woodgate, W., Cayuela, L., Marshall, A. R., Ledo, A., Paine, T. C. E., Marchant, R., Burt, A., Calders, K., Courtney-Mustaphi, C., Cuni-Sanchez, A., Deere, N. J., Denu, D., de Tanago, J. G., Hayward, R., Lau, A., Mac\u0026iacute;a, M. J., Olivier, P. I., \u0026hellip; Platts, P. J. (2018). Tropical forest canopies and their relationships with climate and disturbance: results from a global dataset of consistent field-based measurements. \u003cem\u003eForest Ecosystems\u003c/em\u003e, \u003cem\u003e5\u003c/em\u003e(1), 1\u0026ndash;14. https://doi.org/10.1186/s40663-017-0118-7\u003c/li\u003e\n\u003cli\u003eRanlund, \u0026Aring;., Grafstr\u0026ouml;m, A., Brown, A., Heden\u0026aring;s, H. \u0026amp; Levin, G. 2023. Chapter 4. Designing monitoring systems. In. Allard, A., Keskitalo, C.H. \u0026amp; Brown A. (reds). Monitoring Biodiversity Combining Environmental and Social Data. Routledge. \u003c/li\u003e\n\u003cli\u003eSannier, C., McRoberts, R. E., Fichet, L. V., \u0026amp; Makaga, E. M. K. (2014). Using the regression estimator with landsat data to estimate proportion forest cover and net proportion deforestation in gabon. \u003cem\u003eRemote Sensing of Environment\u003c/em\u003e, \u003cem\u003e151\u003c/em\u003e, 138\u0026ndash;148. https://doi.org/10.1016/j.rse.2013.09.015\u003c/li\u003e\n\u003cli\u003eShirima, D. D. (2015). \u003cem\u003ePhD Thesis: Forests and woodlands of Tanzania: interactions between woody plant structure, diversity, carbon stocks and soil nutrient heterogeneity\u003c/em\u003e [Norwegian University of Life Sciences]. https://nmbu.brage.unit.no/nmbu-xmlui/handle/11250/2579210\u003c/li\u003e\n\u003cli\u003eSileshi, G. W. (2015). The relative standard error as an easy index for checking the reliability of regression coefficients. \u003cem\u003eResearchgate\u003c/em\u003e, \u003cem\u003eDiscussion paper\u003c/em\u003e, \u003cem\u003eAugust\u003c/em\u003e, 1\u0026ndash;25. https://doi.org/10.13140/RG.2.1.2123.6968\u003c/li\u003e\n\u003cli\u003e\u003cu\u003eStehman, S. V., Sohl, T. L., \u0026amp; Loveland, T. R. (2003). Statistical sampling to characterize recent United States land-cover change. Remote Sensing of Environment, 86(4), 517\u0026ndash;529. https://doi.org/10.1016/S0034-4257(03)00129-9\u003c/u\u003e\u003c/li\u003e\n\u003cli\u003eSyampungani, S., Chirwa, P. W., Akinnifesi, F. K., Sileshi, G., \u0026amp; Ajayi, O. C. (2009). The Miombo woodlands at the cross roads: Potential threats, sustainable livelihoods, policy gaps and challenges. \u003cem\u003eNatural Resources Forum\u003c/em\u003e, \u003cem\u003e33\u003c/em\u003e(2), 150\u0026ndash;159. https://doi.org/10.1111/j.1477-8947.2009.01218.x\u003c/li\u003e\n\u003cli\u003eTchatchou, B., Sonwa, D. ., Ifo, S., \u0026amp; Tiani, A. . (2015). \u003cem\u003eDeforestation and forest degradation in the Congo Basin: State of knowledge , current causes and perspectives\u003c/em\u003e (No. 144). https://doi.org/10.17528/cifor/005894\u003c/li\u003e\n\u003cli\u003eURT. (2015). \u003cem\u003eNational Forest Resources Monitoring and Assessment of Tanzania Mainland: Main Results\u003c/em\u003e. United Republic of Tanzania, Ministry of Natural Resources and Tourism. Dar Es Salaam.\u003c/li\u003e\n\u003cli\u003eVerhegghen, A., Kuzelova, K., Syrris, V., \u0026amp; Eva, H. (2022). Mapping Canopy Cover in African Dry Forests from the Combined Use of Sentinel-1 and Sentinel-2 Data : Application to Tanzania for the Year 2018. \u003cem\u003eRemote Sensing\u003c/em\u003e, \u003cem\u003e14\u003c/em\u003e(1522), 1\u0026ndash;21. https://doi.org/https://doi.org/10.3390/rs14061522\u003c/li\u003e\n\u003cli\u003eWestfall, J. A., \u0026amp; Morin, R. S. (2013). A cover-based method to assess forest characteristics using inventory data and GIS. \u003cem\u003eForest Ecology and Management\u003c/em\u003e, \u003cem\u003e298\u003c/em\u003e, 93\u0026ndash;100. https://doi.org/10.1016/j.foreco.2013.02.036\u003c/li\u003e\n\u003cli\u003eZanaga, D., Van De Kerchove, R., De Keersmaecker, W., Souverijns, N., Brockmann, C., Quast, R., Wevers, J., Grosu, A., Paccini, A., Vergnaud, S., Cartus, O., Santoro, M., Fritz, S., Georgieva, I., Lesiv, M., Carter, S., Herold, M., Linlin, L., Tsendbazar, N., \u0026hellip; Arino, O. (2021). ESA WorldCover 10 m 2020 v100. \u003cem\u003eMeteosat Second Generation Evapotranspiration (MET)\u003c/em\u003e.\u003c/li\u003e\n\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":"Two-Stage sampling, Stratified Random Sampling, Auxiliary variables, Horvitz-Thompson estimator, Remote sensing","lastPublishedDoi":"10.21203/rs.3.rs-3880805/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-3880805/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eLong-term monitoring is essential to understand the impacts of land use and climate change on miombo woodlands. This study introduces an innovative monitoring design for miombo woodlands with a two-stage sampling utilizing spatially balanced techniques to estimate the area and canopy cover of miombo woodland across the Tabora, Sikonge, Mlele, and Tanganyika districts. The first step involved the selection of 68 tracts, each comprising an average of 1025 plots, with the aid of spatially balanced sampling. Each of the 69,716 plots was classified into closed (canopy cover\u0026thinsp;\u0026gt;\u0026thinsp;70%), open (40% \u0026le; canopy cover\u0026thinsp;\u0026le;\u0026thinsp;70%), very open (10% \u0026le; canopy cover\u0026thinsp;\u0026lt;\u0026thinsp;40%), and non-miombo (canopy cover\u0026thinsp;\u0026lt;\u0026thinsp;10%) based on woodland cover derived from Sentinel 2 images, followed by the second step consisting of stratified random sampling and inventorying of 2,690 plots within 68 tracts. Using PlanetScope images, we determined the canopy cover for the 2,690 plots selected in the second step and reclassified them accordingly. Employing the Horvitz\u0026ndash;Thompson estimator, our results showed that miombo woodlands in these districts cover 37,359\u0026thinsp;\u0026plusmn;\u0026thinsp;4,618 km\u0026sup2; with an average canopy cover of 55% \u0026plusmn; 5%. Closed miombo woodland (canopy cover\u0026thinsp;\u0026gt;\u0026thinsp;70%) was the dominating woodland type, covering 29,546\u0026thinsp;\u0026plusmn;\u0026thinsp;4,382 km\u0026sup2; of the study area with an average canopy cover of 84% \u0026plusmn; 7%. The study's innovative sampling design provides reliable estimates of the area of miombo woodlands and average canopy cover, with relative standard errors consistently below 25%, offering a robust foundation for monitoring different miombo types.\u003c/p\u003e","manuscriptTitle":"Estimating canopy cover using a Spatially Balanced Sampling approach: A case study of miombo woodlands in Western Tanzania","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2024-02-27 20:58:50","doi":"10.21203/rs.3.rs-3880805/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":"3969cfd1-5675-492f-b3e4-f1ceea5b17c3","owner":[],"postedDate":"February 27th, 2024","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[],"tags":[],"updatedAt":"2024-04-25T13:55:00+00:00","versionOfRecord":[],"versionCreatedAt":"2024-02-27 20:58:50","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-3880805","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-3880805","identity":"rs-3880805","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}