Spatial Patterns of Above-Ground Biomass in Tropical Alpine Páramo Ecosystems Using Allometric Models and LiDAR Data

Spatial Patterns of Above-Ground Biomass in Tropical Alpine Páramo Ecosystems Using Allometric Models and LiDAR Data | 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 Spatial Patterns of Above-Ground Biomass in Tropical Alpine Páramo Ecosystems Using Allometric Models and LiDAR Data Paula Veloza, Anamaría Rozo, Leonardo Segura, Marian Cabrera, and 3 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-5348181/v1 This work is licensed under a CC BY 4.0 License Status: Published Journal Publication published 11 Jul, 2025 Read the published version in Landscape Ecology → Version 1 posted 9 You are reading this latest preprint version Abstract Context Páramos, high-elevation alpine ecosystems found in the northern Andes, are a biodiversity hotspot and play a crucial role in climate change mitigation due to their carbon storage capacity. Above-ground biomass, AGB, serves as a key indicator of ecosystem health and carbon sequestration potential. Accurate estimates of above-ground biomass are essential for understanding the variability of carbon storage across different páramo vegetation types, successional stages and degradation impacts supporting the design of effective conservation and management strategies. Objectives Using a combination of methods from direct measurements to UAS LiDar, we describe the main patterns of above-ground biomass across contrasting vegetation types and plant growth forms in páramos of the northeast Andes of Colombia. Methods This study was conducted in conserved páramo areas in El Cocuy National Natural Park in the northern part of the Colombian Andes. We measured the ABG biomass of the different growth forms and related that to relevant allometric traits by using simple linear models. Using the allometric equations we estimated the AGB of 30 plots in areas dominated by different páramo vegetation types. Airborne LiDAR data was collected from these plots and canopy height and density metrics were processed to determine landscape-level above-ground biomass calibrated with the ground measurements. Results We found that plant height, basal diameter, and leaf area explained above-ground biomass variation for the different growth forms. We selected models with canopy height model (CHM) as predictor, to explain above-ground biomass at the landscape lev el. Allometric and LiDAR derived models showed páramo biomass values ranging from 3 to 11 Mg C ha − 1 . Conclusions Our results demonstrated that it is possible to understand above ground carbon accumulation patterns at the landscape level by combining direct and indirect methods, such as allometric equations and LiDAR data, in areas representing the heterogeneity of páramo vegetation. This study is pioneering in providing information for non-forest carbon reservoirs and the impacts of human actions on the dynamics of the ABG biomass, which are crucial to reach national GHG emission targets. above ground biomass allometry tropical alpine paramo LiDAR carbon dynamics Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Introduction Changes in the dynamics of organic carbon in tropical ecosystems are a key indicator of the effects of climate change and human actions on ecological processes (Chave et al. 2004 ; Gong et al. 2024 ; Heimann and Reichstein 2008 ). Organic carbon plays a crucial role in regulating nutrient cycling, soil health, and biodiversity, while also influencing global carbon budgets. Developing effective strategies to mitigate climate change requires large-scale monitoring and quantification with high resolution and reliable methods (Lu et al. 2016 ). Studies on aboveground biomass estimation in tropical ecosystems have primarily focused on forest (Asner and Mascaro 2014 ; Kerkhoff and Enquist 2009 ; Lu et al. 2016 ; Mascaro et al. 2011 ), whereas non-forest ecosystems such as high Andean grasslands, have significant challenges for accurate carbon quantification (Cabrera and Duivenvoorden 2020 ; Hofstede et al. 1995 ; Oliveras et al. 2014c ). The structural complexity, species diversity, and methodological limitations associated with these ecosystems, has hindered the development of reliable biomass estimation. Páramos are critical ecosystems that provide key ecosystem services, including water and carbon regulation (Farley et al. 2011 ). In these ecosystems, biomass has been estimated at plot level through direct methods (Hofstede et al. 1995 ; Tol and Cleef 1994 ) and indirect methods using allometric equations (Cabrera and Duivenvoorden 2020 ; Minaya et al. 2016 ). Remarkably, methods for detecting changes over larger areas remains unexplored. Addressing the relationship between field-based biomass measurements and remote sensing tools will contribute to a better understanding of páramo carbon stocks and inform effective climate change mitigation strategies. Vegetation scanning methods provide an efficient alternative between small-scale manual measurements and satellite or aerial platforms that cover larger areas (Anderson et al. 2018 ).These tools provide the appropriate temporal and spatial resolution, enabling the capture of the inherent heterogeneity of grasslands (Bazzo et al. 2023 ). Remote sensing methods, such as LiDAR sensors (Light Detection and Ranging), measure structural attributes such as volume, density, or vegetation height, which, together with direct measurements, allow for estimating biomass at landscape-level with high precision (Almeida et al. 2019; Camarretta et al. 2020 ; Tmušić et al. 2020 ; Zhang et al. 2021 ). Unmanned aerial vehicles (UAV) equipped with LiDAR, offer a cost-effective and efficient solution for obtaining precise three-dimensional data over large areas. These sensors can penetrate dense vegetation, providing detailed information on plant structure and height, which is essential for accurate biomass estimation (Sinde-González et al. 2021 ). Unlike passive satellite-based sensors, UAVs avoid saturation issues, ensuring reliable data collection (Almeida et al. 2019; Salas 2021 ; Zhang et al. 2021 ). Plant height and vegetation cover (VC) have been reported as reliable predictors to estimate above ground biomass (Schulze-Brüninghoff et al., 2019; Zhang et al., 2021 ). UAVs methods can be used complementarily to systematically monitor aboveground biomass at a high spatial and temporal resolution. By combining this data with ground-measurements, we can develop accurate and scalable models for estimating aboveground biomass. This approach may significantly enhance the measurement, reporting, and verification of carbon offset schemes, ensuring the effectiveness in mitigating climate change (Yazaki et al. 2016 ). This study aims to estimate aboveground biomass in páramo covers using allometric models and LiDAR in a micro-watershed of the El Cocuy National Natural Park in Colombia. To this end, three specific objectives were addressed: 1) Develop and evaluate non-destructive multi-species allometric models to estimate aboveground biomass in páramo covers. 2) Estimate aboveground biomass in different covers using a LiDAR sensor carried by an unmanned aerial vehicle in combination with allometric models. 3) Quantify the carbon content in aboveground biomass in different páramo ecosystem covers at the landscape level. This article aims to contribute to the understanding of the carbon cycle of the páramo ecosystem. Methods Study site This study was conducted in El Cocuy National Natural Park, located in the northern area of the Eastern Cordillera of Colombia. Our plots were established within the Lagunillas River area, located in the southern sector of the western flank of the Cocuy Sierra Nevada (Muñoz Blanco et al. 2005 ) (Fig. 1 ). The study area is characterized by a bimodal rainfall regime (with rainfall peaks from April to June and from September to November) and annual precipitation ranging between 15.00–20.00 mm. The páramo ecosystem is found at elevations between 3.750 and 4.000m and is represented by vegetation mosaics dominated by Espeletia, Calamagrostis, Pentacalia, Diplsotephium, Linochilus, Ageratina, herbs like Galium and Arenaria and trees of the genus Polylepis (Cleef 1981 ; Ruiz et al. 2008 ). Field Data Our research involved several key steps, including field data collection, statistical modeling, LiDAR data processing, and model selection and evaluation for above-ground biomass estimation from plot to landscape scale (Fig. 2 ). We selected different areas dominated by growth forms such as shrub, caulirosettes, herbs, tussocks, and Polylepis trees. We chose areas with slopes between 20% and 50%, avoiding sites with steep topography, based on the slope map generated from the Digital Elevation Model (NASA SRTM Digital Elevation 30m). In these areas, we established 30 circular plots of 0.07 ha (707 m2) following the methodology of the National Forest Inventory (NFI) for forests in Colombia (Barreto et al., 2018 ), adapted to analyze the heterogeneity of páramo vegetation. The adaptation involved placing a 2x2 m subplot at the center of each circular plot for the measurement of non-woody plants and caulescent rosette species. We followed the growth form classification for páramo vegetation: shrubs, basal rosettes, tussocks, herbs and caulirosettes (Cleef 1981 ; Hughes and Atchison 2015 ; Ramsay and Oxley 2001 ). We measured small shrubs and trees (2.5 cm ≤ DBH < 10 cm) in plots within 3 m radius; medium trees (10 cm ≤ DBH < 30 cm) in subplots within a 7 m radius, and large trees (DBH ≥ 30 cm) within a 15 m radius (Barreto et al. 2018 ). We measure plant height (H, in meters) of all recorded individuals using a Vertex IV Haglöf and measuring tapes (0.1 cm precision). The diameter of each individual was measured using a caliper (0.01 cm precision) and measuring tapes. The circumference was converted to diameter using the formula GBH/π. In the 2x2 m subplots, we measured individuals with growth forms such as herbs, cushions, basal rosettes, caulescent rosettes, and tussock grasses, recording their height, diameter, and foliage area. we measured the diameter at 0.5 cm from the ground, the total height from the ground surface to the apical leaf, the stem height was measured from the ground surface to the base where the rosette begins, and the leaf area. For the caulirosettes, cushions, basal rosettes, tussock grasses, and herbs, the leaf area was measured in the North-South and East-West directions from the leaf tips. To avoid altering the condition of the plots and biomass contents for long-term monitoring, we collected individuals outside the plot at a maximum distance of 5 meters. Finally, we transported the samples to the Ecosystems and Climate Change Laboratory at Universidad Javeriana in Bogotá, where they were stored in individually labeled paper bags and placed in an oven at 60°C for 48 hours until their dry weight stabilized, which was recorded for each individual. Necromass was not included, as it consists of decaying tissue attached to the plant; we included leaves and stems and removed reproductive structures (Körner 2021 ). We collected all vegetation classified as herbs and we collected individuals for all other growth forms. We placed these plants in bags and dried in an oven at 60°C for 48 hours until reaching a constant weight and were subsequently weighed on a precision balance in the Ecosystems and Climate Change laboratory at Pontificia Universidad Javeriana. Allometric models For each growth form, we collected 167 shrubs, 56 cushion plants, 89 stem rosettes, 32 tussock grasses, 63 basal rosettes, and 54 herbs. These individuals were collected in the El Cocuy National Natural Park and other protected páramo areas of the eastern and central Colombian mountain ranges between 2017 and 2024 (Chingaza National Natural Park and Los Nevados National Natural Park). For each growth form, we established different allometric models considering the relationship between biomass and the predictive variables: total height, diameter, stem area, basal cylindrical volume (VCB = π* radius* height) and elliptical cone volume (VCO = 1/3* π *radius2*height) for tussocks and caulirosettes (Johnson et al. 1988 ; McClaran et al. 2013 ) (see appendix 2). We tested these equations using simple power-law functions (Y = a + xb) for each predictor, including the relationship between them. We tested the distribution of each variable to ensure the assumption of normality and homoscedasticity. Variables that deviate from normality were transformed using the logarithm or square root. Model selection for each growth form was based on a combination of visual inspection of residuals, outliers, and influential points for each model, as well as statistical metrics such as AIC weights, mean squared error, and adjusted R² (Zuur et al. 2009 ). We calculated the mean squared error using k-fold cross-validation (Bro et al. 2008 ); in which 60% of the data was selected for model calibration, and the remaining 40% was used for error estimation. These values are reported for each allometric model and growth form. Plot biomass estimation According to the best models selected for each growth form, the equations were applied to the individuals recorded in each plot, except for large tree including Polylepis where used already published equations were used (Alvarez et al. 2012 ; Vásquez et al. 2014 ). Biomass estimates for each individual were back transformed to weight units using the relevant function exponential for those variables log transformed. An additional correction was performed using the correction factor (CF), based on the standard residual sum of squares (RSE²) for each selected model (Baskerville 1972 ; Chave et al. 2005 ). Each plot was analyzed using a Light Detection and Ranging (LiDAR) sensor, specifically the Zenmuse L1 mounted on a DJI Matrice 300 RTK UAV. We selected detailed parameters for the flight mission, including flight pattern, area, altitude, trajectory, speed, and point density, accordingly the conditions of the study sites (Table 1 ). Due to the heterogeneity of the study areas, flight altitudes ranged from 50 to 70 meters, with higher altitudes used in areas where vegetation was taller. Additionally, we conducted flights at lower speeds to obtain dense point clouds, which allowed for the identification of small vegetation such as grasses and herbs (Zhang et al. 2021 ). To ensure comprehensive data coverage, we defined flight areas using polygons that encompassed all plots on the ground for each cover vegetation type, obtaining six flight zones, ranging from 3 to 10 hectares, based on the spatial distribution of ground plots. We conducted an oblique flight mode, capturing data from different angles to obtain detailed three-dimensional point clouds. We georeferenced each plot by locating its centroid using a high-precision RTK antenna and triangulating coordinates obtained from GPS. According to each flight zone, we applied a zonal statistic to the resulting biomass raster at the landscape scale. Table 1 Parameters of flight mission for the different vegetation covers in the Lagunillas paramo area using the Zenmuse L1 sensor on the Matrice 300 RTK UAV. Flight parameter Value Flight parameter Value Flight route mission Oblique Flight zone 3–9 hectares Flight speed 5m/seg Flight duration 15–27 min Average LiDAR Density 950–3.365 points/m 2 GSD (Ground Sample Distance) 1–3 cm/pixel Sensor size 152 x 110 x 169 mm Bands Near Infrared Altitude 50–70 meters Datum WKID 9377 LiDAR data processing We processed LiDAR data using DJI Terra software to reconstruct the raw data into point clouds in .LAS format. We processed point clouds in RStudio software, using integrated algorithms from the LidR and ForestGARP packages (Roussel et al. 2020 ). To generate the Digital Terrain Model (DTM), we classified the ground points through the Triangular Irregular Network (TIN), which uses Delaunay triangulation as a linear interpolation method to produce a more detailed DTM (Roussel et al. 2020 ). With the obtained DTM, we normalized the point cloud to classify the data corresponding to vegetation. Subsequently, we generated the Canopy Height Model (CHM) using the Point to Raster (P2R) algorithm. This algorithm assigns the height of the highest point within its area to each pixel in the resulting raster, applying an adjustment that considers 8 surrounding points to the original. This forms a sub circle to treat each LiDAR point as a disk rather than an individual point. This process densifies the point cloud and produces a more uniform CHM with fewer empty pixels (Roussel et al. 2020 ). Using the CHM and applying the Individual Tree Detection (ITD) algorithm from the LidR package (Roussel et al. 2020 ), we identified individual vegetation by segmenting the canopy. We obtained structural metrics such as maximum canopy height (Zmax), mean canopy height (Zmean), canopy height distribution by percentiles (Zp75), as well as high (Imax), mean (Imean), and percentile-distributed intensity for each segmented individual (Ipcmz10, Ipcmz50, and Ipcmz90). UAV LiDAR metrics We estimated Lorey's Height as the index that calculates the average height of each plant individual in proportion to its basal area for each plot in each cover type. This is a relevant parameter derived from LiDAR data, particularly for ecosystems with heterogeneous structure (Rajab Pourrahmati et al. 2018 ). We obtained the parameters for maximum Lorey height (Zmax), mean Lorey height (Zmean), and the 75th percentile height (Z75) for each plant individual by relating the convex area of the canopy to the heights (Roussel et al. 2020 ). Additionally, we estimated the average height per plot using the Canopy Height Model (CHM). A filter was applied to the segmented vegetation individuals to exclude those composed of fewer than 100 points, with heights below 0.5 m or above 15 m. However, individuals with a higher number of points were retained to represent vegetation shorter than 0.5 m. For each plot within the different cover types, we obtained the Vegetation Cover fraction (VC) (Zhang et al., 2021 ) to derive various metrics for shorter growth forms such as herbs, grasses, basal rosettes, cushion plants, and some caulescent rosettes. We obtained this metric from the CHM and segmented vegetation using the ITD algorithm, producing a factor ranging from 0 to 1. We measured the return intensity, which represents the amount of energy reflected back to the sensor as a function of the illuminated area and canopy reflectance (García et al. 2010 ), according to the equation proposed by Roussel et al. (2024) we obtained the intensity for each segmented plant individual and expressed as the 10th percentile (Ipcmz10), the 50th percentile (Ipcmz50), the 90th percentile (Ipcmz90) and its standard deviation (IpcmzSD). To examine the relationship between metrics obtained from LiDAR and data measured in the field, we performed a Pearson correlation analysis. UAV LiDAR above-ground biomass We estimated above-ground biomass at the plot scale by developing linear regression models, using the above-ground biomass values derived from allometric regressions in each plot as the observed variable and the LiDAR metrics as explanatory variables, while analyzing their collinearity. We tested these equations using simple power law functions (Y = a + xb). We selected the model based on residual standard error, adjusted R² (R² adj), AIC, and visual inspection of the residual plots. These values are reported for each model (see Appendix 2). We evaluated the error in above-ground biomass estimation using the LiDAR sensor by calculating the Root Mean Square Error (RMSE), which provides a quantitative assessment of the performance of the LiDAR models in predicting vegetation biomass (Bazzo et al. 2023 ). A lower RMSE value indicates that the model's predictions are closer to the observed values, suggesting higher accuracy in the estimates and, therefore, better predictive quality in the regression model. Using the equation obtained from the model and the metrics derived from the sensor data processing (CHM_mean), we calculated above-ground biomass at landscape scale, generating a raster with 0.1m pixels. We performed zonal statistics for the total area covered by the UAV LiDAR for each flight zone, averaging the pixel values from the above-ground biomass raster in units of MgCha⁻¹. Pixels corresponding to surface water bodies, such as lakes and lagoons, were assigned NA values, meaning that only vegetation and bare soil were included in the zonal statistics. Results Allometric models The best models based on AIC, R², and the comparison of observed vs. predicted values for the growth forms of caulescent rosettes (R²= 0.52), tussocks (R²= 0.52), and cushions (R²= 0.63) were those using leaf area as a predictor. For herbs (R²= 0.43) and rosettes (R²= 0.76), the best allometric models were those related the product of leaf area and plant height, while for shrubs, the best model contains the product of height and diameter as predictors (R²= 0.60) (Table 2 ). The errors associated with cross-validation for most growth forms were below one gram. Table 2 Selected allometric models for estimating aboveground biomass (AGB) for each growth form of paramo vegetation. Predictors: H (total plant height in cm); FA (leaf area); D (diameter in cm), N (number of plants of the allometric set). Model coefficients: a, b, c, d. AIC: Akaike Information Criterion, R²adj: adjusted R-squared, CV: Error obtained from cross-validation. The average and range of plant height and diameter variables used for calibrating the equations for each growth form in Paramo areas of the Colombian Andes. Growth form Regression model N a b c d AIC R2-adj CV Error Plant Height (McClaran et al.) Diameter (McClaran et al.) Caulirosette Ln(AGB) = a + b*sqrt(AF) 46 3.304 0.075 193 0.52 1.1 86.3 ± 67 (9.6-259.8) 29.23 (5.3–76) Tussocks Ln(AGB) = a + b*sqrt(AF) 19 3.011 0.028 28.63 0.52 0.56 63 ± 5 (33–79) - Shrubs Ln (AGB) = a + b*log(H) + c*log(D) + d*log(H)*log(D) 102 -0.290 0.810 -0.892 0.269 203.5 0.60 0.58 64 ± 4 (18–250 8.5 ± 18 (0.1-11.42 Cushions Ln(AGB) = a + b*log(AF) 32 0.59 0.403 33,40 0.63 0.25 7 ± 10 (1-44.2) Rosettes Ln(AGB) = a + b*sqrt(H)*c*log(AF) + d*sqrt(H)*log(AF) 41 -0.770 0.628 0.113 -0.018 64.8 0.76 0.68 60 ± 55 (1-104) Herbs Ln(AGB) = a + b*sqrt(H) + c*log(AF) + d*sqrt(H)*log(AF) 25 -1.819 0.283 1.086 -0.09 81.23 0.43 0.55 40 ± 14.5 (18–87) 0.5 ± 0.12 (0.3–0.6) UAV LiDAR metrics Based on the individual tree segmentation within the plots, a minimum of 50 individuals were detected for tree and shrub cover, and up to 170 individuals in vegetation dominated by grasses. The relationship between field-measured heights and heights derived from Lorey’s height and the canopy height model (CHM) were very similar (R²= 0.72, 0.73, respectively). With the vegetation cover metric, a greater dispersion of samples and a weaker relationship were observed (R²= 0.24), while the intensity at the 10th percentile showed negative correlations (R²= -0.43), which tended to decrease as the percentile increased. The correlation between aboveground biomass (AGB) obtained through allometric models and height metrics, such as Lorey’s height and the CHM, showed a similar trend to that observed with field-measured heights (R²= 0.65 and R²= 0.66, respectively) (Fig. 3 ) (Appendix 3). Estimation of aboveground biomass using allometry and UAV LiDAR The best model for estimating aboveground biomass at the landscape level using LiDAR is based on the mean canopy height model (mean CHM) and an error around two tons of carbon per hectare (R² = 0.65- RMSE = 2.29). This equation allows calculating the biomass for the different type of covers found in the páramo zones, regardless of their structure (Table 4 ). The biomass expressed through the selected cover types represents the entire heterogeneity of the páramo landscape, where caulirosettes, tussocks, shrub, and trees dominate different areas. Based on the allometric equations, vegetation dominated by shrubs, stored an aboveground biomass of 4.3 ± 2.9 Mg C ha⁻¹. For areas dominated by stem rosettes, contained 3.8 ± 2.0 Mg ha⁻¹ in aboveground biomass. Areas dominates by tussocks and grasses retained 4.3 ± 2 Mg C ha⁻¹ of biomass, while areas dominated by Polylepis trees accumulated around 11.76 ± 7 Mg C ha⁻¹ of biomass. Using LiDAR metrics, vegetation dominated by shrub accumulated on aboveground biomass 5.4 ± 1.3 Mg C ha⁻¹, caulirosette-dominated cover accumulated 3.7 ± 0.5 Mg C ha⁻¹, a value very similar to that reported by the allometric equations. The aboveground biomass accumulated in areas dominated by tussocks, was 3.3 ± 0.4 Mg C ha⁻¹. Finally, areas dominated by Polylepis stored a biomass of 12.1 ± 4.2 Mg C ha⁻¹. These results show a pattern similar to those reported by the allometric equations (Fig. 4 ). Table 4 Model estimators for the estimation of aboveground biomass at the landscape level in the páramo, considering the CHM mean variable (average vegetation cover) obtained with LiDAR. AGB: aboveground biomass, Model coefficients: a, b, c, d. AIC: Akaike Information Criterion, R²adj: adjusted R-squared, RMSE: root mean square error of the model. Cobertura Regression model a b AIC R2-adju RMSE General AGB = a + (b*CHM_mean) 2.56 3.26 140.9 0.65 2.29 Landscape aboveground estimation reflected the vegetation structure of each cover type. In flight zone 1, the average AGB was the highest compared to other areas due to the dominance of Polylepis trees, which have greater height compared to other cover types. Flight zones 3 and 5, predominantly covered by shrubs, showed lower average AGB values compared to zone 1, but higher values when comparing with zones 2, 5 and 6, in which shorter shrubs, grasses, and caulirosettes dominated (Table 5 ). We could determine carbon storage distribution for each flight zone with a very detailed spatial resolution (0.10 m resolution raster) by estimating AGB at landscape scale (Fig. 5 ). Table 5 Aboveground biomass (AGB) estimated at the landscape scale for each cover type in de 6 flight zones based on metrics obtain from LiDAR. We present the flight area covered, the aboveground biomass obtained, and the aboveground biomass estimated at the landscape scale. Flight Area (ha) AGB (Mg C ha − 1 ) AGB land scale (Mg C) Flight Zone1 3.47 8.20 28.45 Flight Zone2 3.43 3.37 11.55 Flight Zone3 9.17 4.45 40.8 Flight Zone4 2.95 3.37 9.94 Flight Zone5 10.58 3.84 40.62 Flight Zone6 5.97 3.27 19.52 Discussion This study is a pioneering in estimating aboveground biomass in tropical high mountain landscapes using allometric models and LiDAR-derived equations. The results suggest that these models have the potential to capture the heterogeneity of vegetation in páramo landscapes, as they accurately predict the aboveground biomass across diverse vegetation structure. Our findings from allometric equations and LiDAR suggest that the tropical páramos of Colombia can store between 3 y 5 Mg C ha − 1 . These values are consistent with values reported in other alpine areas in the Andes with 3.3 Mg C ha − 1 for the grasslands of the Peruvian puna; 4.2 Mg C ha − 1 in Ecuadorian páramos, and the 4 o 5 Mg C ha − 1 in the páramos of the Eastern Cordillera of Colombia (Hofstede et al. 1995 ; Oliveras et al. 2014a ; Ramsay and Oxley 2001 ). This pattern in aboveground biomass in the páramo may be related to factors such as vegetation density and species life span, which directly influence decomposition rates and organic matter production in ecosystems (Pinos et al. 2017 ). Additionally, factors such as changes in soil moisture and nutrients, along with species composition, also play a crucial role in these variations (Cabrera and Duivenvoorden 2020 ; Gao et al. 2021 ). The observed higher values of above-ground biomass in areas dominated by shrub and Polylepis (between 5 and 12 Mg C ha⁻¹), might be associated with the increase in wood production, similar to that has been observed in the páramos of the Eastern range of Colombia (Cardozo and Schnetter 1976 ). These values fall within the range reported for shrublands in the Himalayas, an alpine vegetation type similar to the tropical páramo, with 6 Mg C ha⁻¹ (1447,31 g m⁻²) aboveground biomass (Nie et al. 2018 ). In contrast, our biomass estimates for Polylepis were lower than those reported in Ecuador (14 and 200 Mg C ha⁻¹) (Montalvo et al. 2018 ). This discrepancy may be attributed to differences in environmental conditions such as altitude, climate and soil characteristics. Moreover, we observed a lower density of Polylepis individuals per hectare than the ones observed in Ecuador (Montalvo et al. 2018 ), with an average of 48 individuals/hectare compared to over 100 individuals/hectare respectively. Understanding the factors driving this variation might be a further step, which is crucial for accurately assessing carbon sequestration potential in the páramo region. To improve the accuracy of biomass estimates, we developed multi-species allometric models for each growth form separately. These equations incorporated the variation in morphological characteristics and biomass accumulation, reflecting the different carbon assimilation mechanisms that determine each growth form (Dorrepaal 2007 ). For most growth forms, allometric models relied on predictors such as height and leaf area (Tol and Cleef 1994 ). Similarly, for tussocks, leaf area was primary predictor of biomass, consistent with findings from other Andean grasslands studies (Table 2 ) (McClaran et al. 2013 ; Rojo et al. 2017 ). For shrubs, the best models included height and stem diameter as predictors, similar to the allometric models used for tropical trees (Alvarez et al. 2012 ; Chave et al. 2005 ) and shrubs, where the combination of height and diameter improves model estimates. In contrast with other studies that use crown diameter (Grigal and Ohmann 1977 ; Smith and Brand 1983 ; Torres et al. 2012 ). This may be due to many of the collected species having few branches, such as some shrub species in the genus Monticalia . For herbs, although the R 2 is lower than reported in other studies, the error is below 1 g. The low R 2 may be explained by the small sample size used to calibrate the allometric model, which can contribute to increase model uncertainty, as a reduced number of samples decreases the predictive power of the models. Therefore, it is recommended to use a larger number of individuals (80–100) (Chave et al. 2005 ; Sileshi 2014 ). This is a challenge for some growth forms where densities are low. Nevertheless, providing information for estimating biomass in Andean vegetation is essential for medium- and long-term carbon monitoring. By using growth form-specific allometric models, we were able to capture the variation in biomass-physiognomy relationship and obtain more reliable estimates of carbon storage in páramo ecosystems. Structural variables of páramo vegetation detected through LiDAR showed a correlation with parameters measured in ground-plots, confirming findings from previous research that highlight the usefulness of three-dimensional LiDAR models for detecting vegetation variability, including small sized species such as grasses and herbs (Wang et al. 2017 ; Zhang et al. 2017 ). In this study, a correlation of 0.73 and 0.72 was observed between field-measured height and height derived from the Canopy Height Model (CHM) and Lorey’s Height, underscoring the importance of height as a key structural variable for estimating aerial biomass (AGB) in grassland ecosystems (Bazzo et al. 2023 ). Additionally, our results were consistent with those reported by Zhang et al. ( 2021 ) where LiDAR metrics showed a strong correlation with biomass obtained through allometry, height, and FVC. The individual segmentation of smaller vegetation remains a challenge, as it depends on parameters established during data acquisition with the sensor, such as flight speed, flight shape, and altitude, as well as the performance of the algorithm used during data processing. To address this challenge, our study planned flights at lower speeds and altitudes, which allowed for the acquisition of dense point clouds that facilitated the identification of low vegetation, such as grasses, herbs, and cushions plants (Zhang et al. 2021 ). Additionally, algorithm parameters were determined to improve individual segmentation (Roussel et al. 2020 ). The structural complexity of grasslands, along with height and species richness, has been identified as an influencing factor in some spectral properties of vegetation (Villoslada et al. 2020 ). To tackle this, we produced several metrics from LiDAR data, focusing on the structural diversity and heterogeneity of páramo vegetation. The selected model was based on predictors derived from the mean plot-level heights obtained from the Canopy Height Model (CHM), which have proven effective across a wide range of grasslands globally (Kümmerer et al. 2023 ; Wang et al. 2017 ; Zhang et al. 2021 ; Zhang et al. 2017 ). This approach allowed for the integration of different growth forms present in all the studied coverages, successfully predicting aerial biomass for the heterogeneity of the páramo at a landscape scale with a coefficient of determination (R²= 0.67) and an error of 2 Mg C ha⁻¹. These results are comparable to those obtained by Chan et al. ( 2021 ) in subtropical forests and exceed those reported by Zhang et al. ( 2021 ) in Mongolian grasslands. As reported by González-Jaramillo et al. ( 2018 ) regarding biomass prediction using LiDAR in grasses and caulirosette-dominated coverages, certain limitations were evident related, showing an underestimation of above-ground biomass. For this vegetation, the segmentation of the vegetation structure is not as accurate compared to the segmentation of trees and shrubs, where their structure is well-defined by three-dimensional models. Particularly for caulirosettes, the variables in the watershed segmentation algorithm had limitations in identifying individual’s due to their morphology (Peitzsch et al. 2024 ). This might be related to the structure of the trunk, which is composed of dead leaves retained in form of necromass and for some other individuals, it may be related to a shorter or not fully developed stems. Their detection may require alternative algorithms that incorporate variables adapted to the structure of this type of vegetation. Although these results are of great importance for proposing innovative methodologies for monitoring the carbon cycle of high mountain ecosystems in the Andes, allowing the evaluation of their status and in some cases, the efficiency of implemented passive or active restoration measures, the comparison of biomass estimates in different studies still presents challenges due to the lack of uniformity in sampling methodologies (Oliveras et al. 2014b ). Therefore, further studies with standardized methodologies are recommended to improve comparability and understanding of these ecosystems. It is still necessary to continue developing new models that integrate metrics obtained from complementary UAV sensors with allometric models, mainly for these ecosystems that present such heterogeneous vegetation. The allometric and LiDAR based models presented here can be applied in remote tropical alpine ecosystems where often the topography of the Andes can be a limiting factor, and remote sensing strategies that circumvent the logistical constraints, the difficulties of accessing remote locations and allows precise ABG carbon estimates in remote landscapes such as the Andean páramos. Conclusions Our results confirm that UVA can be used to estimate the biomass of high mountain grasslands and shrublands; however the method could be improved for rosettes, specifically for species of the genus Espeletia . Additionally, the great potential to estimate biomass at a regional scale with a high degree of accuracy allows for a more accurate estimation of above ground biomass and carbon accumulation patterns on an endangered ecosystem. Mos important, the methods open the possibility of monitoring large areas of paramo areas at low cost and high precision including restored, burned or grazed areas, offering an alternative to the different conservation agencies currently engaged in the protection of the páramo. Declarations Author Contribution PV: Formal analysis, research, visualization; Writing: original draft preparation and review and editing; data collection. AR: Formal analysis; research; visualization; Writing: original draft preparation and review and editing; data collection. LS: Formal analysis; research; visualization; Writing: original draft preparation, review, and editing. MS: Formal analysis; research; Writing: review and editing.JCB: Conceptualization; funding acquisition; supervision; Writing: review and editing. LFP: Conceptualization; funding acquisition; edititing. FN: Conceptualization; funding acquisition; edititing. Acknowledgement We would like to thank the Pontificia Universidad Javeriana of Bogotá for the opportunity to research this high mountain topic and for the use of its facilities. To Ecopetrol S.A and the ICEPT (Instituto Colombiano del Petróleo y Las Energías de la Transición) for its funding and support of the research. To the National Natural Park El Cocuy and all the park rangers for their support in the field. To Marcos Correa, Edward Avilán, Laura Báez, Silvio Ortegón, Edna Rincón, Angie Gómez, Freddy Avellaneda, Astrid Cruz, and David Hernández for their support in the field and laboratory activities. Data availability The authors declare that the data supporting the findings of this study are available within the paper and its Supplementary Information files. Should any raw data files be needed in another format they are available from the corresponding author upon reasonable request. References Almeida DRAd, Stark SC, Shao G et al (2019) Optimizing the remote detection of tropical rainforest structure with airborne lidar: Leaf area profile sensitivity to pulse density and spatial sampling. Remote Sensing 11(1):92 Alvarez E, Duque A, Saldarriaga J et al (2012) Tree above-ground biomass allometries for carbon stocks estimation in the natural forests of Colombia. Forest Ecology and management 267:297-308 Anderson KE, Glenn NF, Spaete LP et al (2018) Estimating vegetation biomass and cover across large plots in shrub and grass dominated drylands using terrestrial lidar and machine learning. Ecological Indicators 84:793-802 Asner GP, Mascaro J (2014) Mapping tropical forest carbon: Calibrating plot estimates to a simple LiDAR metric. Remote Sensing of Environment 140:614-624 Barreto J, Capachero C, Barbosa A, Phillips J, Peña M (2018) Manual de Campo Inventario Forestal Nacional de Colombia. Baskerville G (1972) Use of logarithmic regression in the estimation of plant biomass. Canadian Journal of Forest Research 2(1):49-53 Bazzo COG, Kamali B, Hütt C, Bareth G, Gaiser T (2023) A review of estimation methods for aboveground biomass in grasslands using UAV. Remote Sensing 15(3):639 Bro R, Kjeldahl K, Smilde AK, Kiers H (2008) Cross-validation of component models: a critical look at current methods. Analytical and bioanalytical chemistry 390:1241-1251 Cabrera M, Duivenvoorden JF (2020) Drivers of aboveground biomass of high mountain vegetation in the Andes. Acta Oecologica 102:103504 Camarretta N, Harrison PA, Bailey T et al (2020) Monitoring forest structure to guide adaptive management of forest restoration: a review of remote sensing approaches. New Forests 51(4):573-596 Cardozo H, Schnetter M-L (1976) Estudios ecológicos en el Páramo de Cruz Verde, Colombia. III. La biomasa de tres asociaciones vegetales y la productividad de Calamagrostis effusa (HBK) Steud. y Paepalanthus columbiensis Ruhl. en comparación con la concentración de clorofila. Caldasia:69-83 Chan E, Fung T, Wong F (2021) Estimating above-ground biomass of subtropical forest using airborne LiDAR in Hong Kong, Sci. Rep.-UK, 11, 1751. Chave J, Andalo C, Brown S et al (2005) Tree allometry and improved estimation of carbon stocks and balance in tropical forests. Oecologia 145:87-99 Chave J, Condit R, Aguilar S, Hernandez A, Lao S, Perez R (2004) Error propagation and scaling for tropical forest biomass estimates. Philosophical Transactions of the Royal Society of London. Series B: Biological Sciences 359(1443):409-420 Cleef A (1981) The vegetation of the páramos of the Colombian Cordillera Oriental. Mededelingen van het Botanisch Museum en Herbarium van de Rijksuniversiteit te Utrecht 481(1):1-320 Dorrepaal E (2007) Are plant growth‐form‐based classifications useful in predicting northern ecosystem carbon cycling feedbacks to climate change? Journal of Ecology 95(6):1167-1180 Farley KA, Anderson WG, Bremer LL, Harden CP (2011) Compensation for ecosystem services: an evaluation of efforts to achieve conservation and development in Ecuadorian páramo grasslands. Environmental Conservation 38(4):393-405 Gao X, Dong S, Xu Y et al (2021) Plant biomass allocation and driving factors of grassland revegetation in a Qinghai‐Tibetan Plateau chronosequence. Land Degradation & Development 32(4):1732-1741 García M, Riaño D, Chuvieco E, Danson FM (2010) Estimating biomass carbon stocks for a Mediterranean forest in central Spain using LiDAR height and intensity data. Remote Sensing of Environment 114(4):816-830 Gong K, Huang Z, Qu M et al (2024) Influences of climate change on carbon and water fluxes of the ecosystem in the Qinling Mountains of China. Ecological Indicators 166:112504 González-Jaramillo V, Fries A, Zeilinger J, Homeier J, Paladines-Benitez J, Bendix J (2018) Estimation of above ground biomass in a tropical mountain forest in southern Ecuador using airborne LiDAR data. Remote Sensing 10(5):660 Grigal D, Ohmann L (1977) Biomass estimation for some shrubs from northeastern Minnesota. North Central Forest Experiment Station, Forest Service. US Department of Agriculture, Research Note, NC-226 3 Heimann M, Reichstein M (2008) Terrestrial ecosystem carbon dynamics and climate feedbacks. Nature 451(7176):289-292 Hofstede RG, Castillo MXM, Osorio CMR (1995) Biomass of grazed, burned, and undisturbed páramo grasslands, Colombia. I. Aboveground vegetation. Arctic and Alpine Research 27(1):1-12 Hughes CE, Atchison GW (2015) The ubiquity of alpine plant radiations: from the Andes to the Hengduan Mountains. New Phytologist 207(2):275-282 Johnson PS, Johnson CL, West NE (1988) Estimation of phytomass for ungrazed crested wheatgrass plants using allometric equations. Rangeland Ecology & Management/Journal of Range Management Archives 41(5):421-425 Kerkhoff AJ, Enquist BJ (2009) Multiplicative by nature: why logarithmic transformation is necessary in allometry. Journal of Theoretical Biology 257(3):519-521 Körner C (2021) The alpine life zone. Alpine plant life: Functional plant ecology of high mountain ecosystems:23-51 Kümmerer R, Noack PO, Bauer B (2023) Using high-resolution UAV imaging to measure canopy height of diverse cover crops and predict biomass. Remote Sensing 15(6):1520 Lu D, Chen Q, Wang G, Liu L, Li G, Moran E (2016) A survey of remote sensing-based aboveground biomass estimation methods in forest ecosystems. International Journal of Digital Earth 9(1):63-105 Mascaro J, Detto M, Asner GP, Muller-Landau HC (2011) Evaluating uncertainty in mapping forest carbon with airborne LiDAR. Remote Sensing of Environment 115(12):3770-3774 McClaran M, McMurtry C, Archer S (2013) A tool for estimating impacts of woody encroachment in arid grasslands: allometric equations for biomass, carbon and nitrogen content in Prosopis velutina. Journal of Arid Environments 88:39-42 Minaya V, Corzo G, Romero-Saltos H et al (2016) Altitudinal analysis of carbon stocks in the Antisana páramo, Ecuadorian Andes. Journal of Plant Ecology 9(5):553-563 Montalvo J, Minga D, Verdugo A et al (2018) Características morfológico-funcionales, diversidad arbórea, tasa de crecimiento y de secuestro de carbono en especies y ecosistemas de Polylepis del sur de Ecuador. Ecología austral 28(1):249-261 Muñoz Blanco FU, Ariano R, Buitrago VR, Reyes Leal M, Carvajal R (2005) Plan de Manejo Ambiental Parque Nacional Natural El Cocuy. Parques Nacionales Naturales de Colombia, Bogota, pp. 114 Nie X-q, Yang L-c, Xiong F, Li C-b, Fan L, Zhou G-y (2018) Aboveground biomass of the alpine shrub ecosystems in Three-River Source Region of the Tibetan Plateau. Journal of Mountain Science 15(2):357-363 Oliveras I, Eynden M, Malhi Y et al (2014a) Grass allometry and estimation of above‐ground biomass in tropical alpine tussock grasslands. Austral ecology 39(4):408-415 Oliveras I, Girardin C, Doughty CE et al (2014b) Andean grasslands are as productive as tropical cloud forests. Environmental Research Letters 9(11):115011 Oliveras I, van der Eynden M, Malhi Y et al (2014c) Grass allometry and estimation of above‐ground biomass in tropical alpine tussock grasslands. Austral ecology 39(4):408-415 Peitzsch EH, Martin-Mikle C, Hendrikx J, Birkeland K, Fagre D (2024) Characterizing vegetation and return periods in avalanche paths using lidar and aerial imagery. Arctic, Antarctic, and Alpine Research 56(1):2310333 Pinos J, Studholme A, Carabajo A, Gracia C (2017) Leaf Litterfall and Decomposition of Polylepis reticulata in the Treeline of the Ecuadorian Andes. Mountain Research and Development 37(1):87-96 Rajab Pourrahmati M, Baghdadi N, Darvishsefat AA et al (2018) Mapping Lorey’s height over Hyrcanian forests of Iran using synergy of ICESat/GLAS and optical images. European Journal of Remote Sensing 51(1):100-115 Ramsay PM, Oxley ERB (2001) An assessment of aboveground net primary productivity in Andean grasslands of central Ecuador. Mountain Research and Development 21(2):161-167 Rojo V, Arzamendia Y, Pérez C, Baldo J, Vilá B (2017) Double sampling methods in biomass estimates of Andean shrubs and tussocks. Rangeland Ecology & Management 70(6):718-722 Roussel J-R, Auty D, Coops NC et al (2020) lidR: An R package for analysis of Airborne Laser Scanning (ALS) data. Remote Sensing of Environment 251:112061 Ruiz D, Moreno HA, Gutiérrez ME, Zapata PA (2008) Changing climate and endangered high mountain ecosystems in Colombia. Science of the total environment 398(1-3):122-132 Salas EAL (2021) Waveform LiDAR concepts and applications for potential vegetation phenology monitoring and modeling: a comprehensive review. Geo-Spatial Information Science 24(2):179-200 Sileshi GW (2014) A critical review of forest biomass estimation models, common mistakes and corrective measures. Forest Ecology and Management 329:237-254 Sinde-González I, Gil-Docampo M, Arza-García M et al (2021) Biomass estimation of pasture plots with multitemporal UAV-based photogrammetric surveys. International Journal of Applied Earth Observation and Geoinformation 101:102355 Smith WB, Brand GJ (1983) Allometric biomass equations for 98 species of herbs, shrubs, and small trees. Research Note NC-299. St. Paul, MN: US Dept. of Agriculture, Forest Service, North Central Forest Experiment Station. Tmušić G, Manfreda S, Aasen H et al (2020) Current practices in UAS-based environmental monitoring. Remote Sensing 12(6):1001 Tol GJ, Cleef AM (1994) Above-ground biomass structure of a Chusquea tessellata bamboo páramo, Chingaza National Park, Cordillera Oriental, Colombia. Vegetatio 115:29-39 Torres AM, Peña EJ, Zúñiga O, Peña JA (2012) Evaluación del impacto de actividades antrópicas en el almacenamiento de carbono en biomasa vegetal en ecosistemas de alta montaña de Colombia. Boletín Científico. Centro de Museos. Museo de Historia Natural 16(1):132-142 Vásquez E, Ladd B, Borchard N (2014) Carbon storage in a high-altitude Polylepis woodland in the Peruvian Andes. Alpine botany 124:71-75 Villoslada M, Bergamo TF, Ward R et al (2020) Fine scale plant community assessment in coastal meadows using UAV based multispectral data. Ecological indicators 111:105979 Wang D, Xin X, Shao Q, Brolly M, Zhu Z, Chen J (2017) Modeling aboveground biomass in Hulunber grassland ecosystem by using unmanned aerial vehicle discrete lidar. Sensors 17(1):180 Yazaki T, Hirano T, Sano T (2016) Biomass accumulation and net primary production during the early stage of secondary succession after a severe forest disturbance in northern Japan. Forests 7(11):287 Zhang X, Bao Y, Wang D et al (2021) Using uav lidar to extract vegetation parameters of inner mongolian grassland. Remote Sensing 13(4):656 Zhang X, He G, Zhang Z, Peng Y, Long T (2017) Spectral-spatial multi-feature classification of remote sensing big data based on a random forest classifier for land cover mapping. Cluster Computing 20:2311-2321 Zuur AF, Ieno EN, Walker NJ, Saveliev AA, Smith GM (2009) Mixed effects models and extensions in ecology with R. Springer Additional Declarations No competing interests reported. Supplementary Files paramoallometrySI.docx Cite Share Download PDF Status: Published Journal Publication published 11 Jul, 2025 Read the published version in Landscape Ecology → Version 1 posted Editorial decision: Revision requested 26 Jan, 2025 Reviews received at journal 15 Jan, 2025 Reviews received at journal 23 Dec, 2024 Reviewers agreed at journal 17 Dec, 2024 Reviewers agreed at journal 07 Nov, 2024 Reviewers invited by journal 05 Nov, 2024 Editor assigned by journal 05 Nov, 2024 Submission checks completed at journal 29 Oct, 2024 First submitted to journal 28 Oct, 2024 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-5348181","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":375314213,"identity":"4325a2a4-06eb-4d1a-9aea-6c4b0fdd47e2","order_by":0,"name":"Paula Veloza","email":"","orcid":"","institution":"Pontificia Universidad Javeriana","correspondingAuthor":false,"prefix":"","firstName":"Paula","middleName":"","lastName":"Veloza","suffix":""},{"id":375314214,"identity":"6931b9a4-1293-415c-8361-9f4d8e4c9043","order_by":1,"name":"Anamaría Rozo","email":"","orcid":"","institution":"Pontificia Universidad Javeriana","correspondingAuthor":false,"prefix":"","firstName":"Anamaría","middleName":"","lastName":"Rozo","suffix":""},{"id":375314215,"identity":"b9145f1f-cfbc-4216-abc0-2ac188067d6f","order_by":2,"name":"Leonardo Segura","email":"","orcid":"","institution":"Pontificia Universidad Javeriana","correspondingAuthor":false,"prefix":"","firstName":"Leonardo","middleName":"","lastName":"Segura","suffix":""},{"id":375314216,"identity":"8aacd84f-3412-45d8-98b1-642307e4368b","order_by":3,"name":"Marian Cabrera","email":"","orcid":"","institution":"Pontificia Universidad Javeriana","correspondingAuthor":false,"prefix":"","firstName":"Marian","middleName":"","lastName":"Cabrera","suffix":""},{"id":375314217,"identity":"813d730f-641a-45ad-bb7c-7a88ae966e8d","order_by":4,"name":"Freddy Niño","email":"","orcid":"","institution":"Instituto Colombiano del Petróleo y Las Energías de la Transición-ICEPT","correspondingAuthor":false,"prefix":"","firstName":"Freddy","middleName":"","lastName":"Niño","suffix":""},{"id":375314218,"identity":"2b27f8da-fbb7-4176-8600-739ae5aef7fa","order_by":5,"name":"Luis Fernando Prado-Castillo","email":"","orcid":"","institution":"Instituto Colombiano del Petróleo y Las Energías de la Transición-ICEPT","correspondingAuthor":false,"prefix":"","firstName":"Luis","middleName":"Fernando","lastName":"Prado-Castillo","suffix":""},{"id":375314219,"identity":"6fb5dff0-7e19-459a-8f2d-2cdc215c0a77","order_by":6,"name":"Juan C. Benavides","email":"data:image/png;base64,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","orcid":"","institution":"Pontificia Universidad Javeriana","correspondingAuthor":true,"prefix":"","firstName":"Juan","middleName":"C.","lastName":"Benavides","suffix":""}],"badges":[],"createdAt":"2024-10-28 15:08:21","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-5348181/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-5348181/v1","draftVersion":[],"editorialEvents":[{"content":"https://doi.org/10.1007/s10980-025-02159-0","type":"published","date":"2025-07-11T15:57:41+00:00"}],"editorialNote":"","failedWorkflow":false,"files":[{"id":68506413,"identity":"8ce6814b-35ac-4fb2-b1f7-742cb7de61c4","added_by":"auto","created_at":"2024-11-08 04:37:22","extension":"jpeg","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":97381,"visible":true,"origin":"","legend":"\u003cp\u003eLocation of the plots within the Lagunillas micro-watershed, El Cocuy National Natural Park (A), El Cocuy Natural National Park located in the Eastern Cordillera (B) Páramo ecosystem in the study area (C) Distribution of plots according to vegetation type.\u003c/p\u003e","description":"","filename":"floatimage1.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-5348181/v1/ad6f01522a5898442880348f.jpeg"},{"id":68506409,"identity":"6b4f18ff-8f41-4ac7-bb2f-0120eafbfd3d","added_by":"auto","created_at":"2024-11-08 04:37:22","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":107843,"visible":true,"origin":"","legend":"\u003cp\u003eData processing to estimate aboveground biomass across different páramo cover types using allometric methods and LiDAR. AGB: aboveground biomass.\u003c/p\u003e","description":"","filename":"floatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-5348181/v1/c01f7cade760c0c399ea4ed2.png"},{"id":68506412,"identity":"1b9d5240-aa2a-47a9-a8d1-e3bf89356e41","added_by":"auto","created_at":"2024-11-08 04:37:22","extension":"jpeg","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":188404,"visible":true,"origin":"","legend":"\u003cp\u003eRelationship between the average of Height (m) metrics, factor vegetation cover (%), and intensity (%) obtained through LiDAR sensors for each plot, related to the biomass obtained through allometry and the heights measured in the field in the 30 plots in Lagunillas páramo, PNN El Cocuy.\u003c/p\u003e","description":"","filename":"floatimage3.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-5348181/v1/b5f836cc2e86b679f40a7840.jpeg"},{"id":68506414,"identity":"87278514-560b-4291-8c6b-1a85953df8d9","added_by":"auto","created_at":"2024-11-08 04:37:22","extension":"jpeg","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":99166,"visible":true,"origin":"","legend":"\u003cp\u003e(A) Aboveground biomass (AGB) estimated using allometric equations for each cover type analyzed in the páramo ecosystem; (B) Aboveground biomass estimated with LiDAR for each cover type (Herbs, Caulirosettes, Shrubs, Polylepis trees); (C) Linear regression between aboveground biomass measured with allometric equations and that measured with LiDAR for the 30 study plots.\u003c/p\u003e","description":"","filename":"floatimage4.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-5348181/v1/6d3835819e1bf56fbd6d4442.jpeg"},{"id":68506912,"identity":"58c8342a-0a8a-40e4-9e77-481d95782708","added_by":"auto","created_at":"2024-11-08 04:45:22","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":732543,"visible":true,"origin":"","legend":"\u003cp\u003eBiomass estimated at the landscape scale in areas with different type of cover in the páramo ecosystem. Left panel: aboveground biomass in Mg C ha⁻¹ at the landscape scale. Right panel: LiDAR point cloud in natural color (RGB). (A) Flight area 1, where Polylepis trees are observed as the dominant vegetation on the right. (B) Flight area 2, primarily composed of smaller shrubs, herbs, and caulirosettes. (C) Flight area 3, dominated by shrubs and caulirosettes. (D) Flight area 4, dominated by caulirosettes. (E) Flight area 5, dominated by smaller shrubs, herbs, and caulirosettes. (F) Flight area 6, composed of herbs and grasses.\u003c/p\u003e","description":"","filename":"floatimage5.png","url":"https://assets-eu.researchsquare.com/files/rs-5348181/v1/207cc3741eaf93a5dd59efa0.png"},{"id":86699431,"identity":"956c115e-a901-4d08-bc51-f1601831639d","added_by":"auto","created_at":"2025-07-14 16:09:34","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":1927634,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-5348181/v1/0d11478b-6e1e-4b4d-97ef-db071cd1c1e8.pdf"},{"id":68506913,"identity":"00376567-5be3-4589-8b61-451bc13c650a","added_by":"auto","created_at":"2024-11-08 04:45:22","extension":"docx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":29222,"visible":true,"origin":"","legend":"","description":"","filename":"paramoallometrySI.docx","url":"https://assets-eu.researchsquare.com/files/rs-5348181/v1/c71ee1552a675e6f102f1a3f.docx"}],"financialInterests":"No competing interests reported.","formattedTitle":"Spatial Patterns of Above-Ground Biomass in Tropical Alpine Páramo Ecosystems Using Allometric Models and LiDAR Data","fulltext":[{"header":"Introduction","content":"\u003cp\u003eChanges in the dynamics of organic carbon in tropical ecosystems are a key indicator of the effects of climate change and human actions on ecological processes (Chave et al. \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e2004\u003c/span\u003e; Gong et al. \u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e2024\u003c/span\u003e; Heimann and Reichstein \u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e2008\u003c/span\u003e). Organic carbon plays a crucial role in regulating nutrient cycling, soil health, and biodiversity, while also influencing global carbon budgets. Developing effective strategies to mitigate climate change requires large-scale monitoring and quantification with high resolution and reliable methods (Lu et al. \u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e2016\u003c/span\u003e). Studies on aboveground biomass estimation in tropical ecosystems have primarily focused on forest (Asner and Mascaro \u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e2014\u003c/span\u003e; Kerkhoff and Enquist \u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e2009\u003c/span\u003e; Lu et al. \u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e2016\u003c/span\u003e; Mascaro et al. \u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e2011\u003c/span\u003e), whereas non-forest ecosystems such as high Andean grasslands, have significant challenges for accurate carbon quantification (Cabrera and Duivenvoorden \u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e2020\u003c/span\u003e; Hofstede et al. \u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e1995\u003c/span\u003e; Oliveras et al. \u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e2014c\u003c/span\u003e). The structural complexity, species diversity, and methodological limitations associated with these ecosystems, has hindered the development of reliable biomass estimation. P\u0026aacute;ramos are critical ecosystems that provide key ecosystem services, including water and carbon regulation (Farley et al. \u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e2011\u003c/span\u003e). In these ecosystems, biomass has been estimated at plot level through direct methods (Hofstede et al. \u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e1995\u003c/span\u003e; Tol and Cleef \u003cspan citationid=\"CR52\" class=\"CitationRef\"\u003e1994\u003c/span\u003e) and indirect methods using allometric equations (Cabrera and Duivenvoorden \u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e2020\u003c/span\u003e; Minaya et al. \u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e2016\u003c/span\u003e). Remarkably, methods for detecting changes over larger areas remains unexplored. Addressing the relationship between field-based biomass measurements and remote sensing tools will contribute to a better understanding of p\u0026aacute;ramo carbon stocks and inform effective climate change mitigation strategies.\u003c/p\u003e \u003cp\u003eVegetation scanning methods provide an efficient alternative between small-scale manual measurements and satellite or aerial platforms that cover larger areas (Anderson et al. \u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e2018\u003c/span\u003e).These tools provide the appropriate temporal and spatial resolution, enabling the capture of the inherent heterogeneity of grasslands (Bazzo et al. \u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). Remote sensing methods, such as LiDAR sensors (Light Detection and Ranging), measure structural attributes such as volume, density, or vegetation height, which, together with direct measurements, allow for estimating biomass at landscape-level with high precision (Almeida et al. 2019; Camarretta et al. \u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e2020\u003c/span\u003e; Tmušić et al. \u003cspan citationid=\"CR51\" class=\"CitationRef\"\u003e2020\u003c/span\u003e; Zhang et al. \u003cspan citationid=\"CR58\" class=\"CitationRef\"\u003e2021\u003c/span\u003e). Unmanned aerial vehicles (UAV) equipped with LiDAR, offer a cost-effective and efficient solution for obtaining precise three-dimensional data over large areas. These sensors can penetrate dense vegetation, providing detailed information on plant structure and height, which is essential for accurate biomass estimation (Sinde-Gonz\u0026aacute;lez et al. \u003cspan citationid=\"CR49\" class=\"CitationRef\"\u003e2021\u003c/span\u003e). Unlike passive satellite-based sensors, UAVs avoid saturation issues, ensuring reliable data collection (Almeida et al. 2019; Salas \u003cspan citationid=\"CR47\" class=\"CitationRef\"\u003e2021\u003c/span\u003e; Zhang et al. \u003cspan citationid=\"CR58\" class=\"CitationRef\"\u003e2021\u003c/span\u003e). Plant height and vegetation cover (VC) have been reported as reliable predictors to estimate above ground biomass (Schulze-Br\u0026uuml;ninghoff et al., 2019; Zhang et al., \u003cspan citationid=\"CR58\" class=\"CitationRef\"\u003e2021\u003c/span\u003e). UAVs methods can be used complementarily to systematically monitor aboveground biomass at a high spatial and temporal resolution. By combining this data with ground-measurements, we can develop accurate and scalable models for estimating aboveground biomass. This approach may significantly enhance the measurement, reporting, and verification of carbon offset schemes, ensuring the effectiveness in mitigating climate change (Yazaki et al. \u003cspan citationid=\"CR57\" class=\"CitationRef\"\u003e2016\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eThis study aims to estimate aboveground biomass in p\u0026aacute;ramo covers using allometric models and LiDAR in a micro-watershed of the El Cocuy National Natural Park in Colombia. To this end, three specific objectives were addressed: 1) Develop and evaluate non-destructive multi-species allometric models to estimate aboveground biomass in p\u0026aacute;ramo covers. 2) Estimate aboveground biomass in different covers using a LiDAR sensor carried by an unmanned aerial vehicle in combination with allometric models. 3) Quantify the carbon content in aboveground biomass in different p\u0026aacute;ramo ecosystem covers at the landscape level. This article aims to contribute to the understanding of the carbon cycle of the p\u0026aacute;ramo ecosystem.\u003c/p\u003e"},{"header":"Methods","content":"\u003cp\u003eStudy site\u003c/p\u003e \u003cp\u003eThis study was conducted in El Cocuy National Natural Park, located in the northern area of the Eastern Cordillera of Colombia. Our plots were established within the Lagunillas River area, located in the southern sector of the western flank of the Cocuy Sierra Nevada (Mu\u0026ntilde;oz Blanco et al. \u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e2005\u003c/span\u003e) (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e). The study area is characterized by a bimodal rainfall regime (with rainfall peaks from April to June and from September to November) and annual precipitation ranging between 15.00\u0026ndash;20.00 mm. The p\u0026aacute;ramo ecosystem is found at elevations between 3.750 and 4.000m and is represented by vegetation mosaics dominated by Espeletia, Calamagrostis, Pentacalia, Diplsotephium, Linochilus, Ageratina, herbs like Galium and Arenaria and trees of the genus Polylepis (Cleef \u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e1981\u003c/span\u003e; Ruiz et al. \u003cspan citationid=\"CR46\" class=\"CitationRef\"\u003e2008\u003c/span\u003e).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eField Data\u003c/p\u003e \u003cp\u003eOur research involved several key steps, including field data collection, statistical modeling, LiDAR data processing, and model selection and evaluation for above-ground biomass estimation from plot to landscape scale (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e). We selected different areas dominated by growth forms such as shrub, caulirosettes, herbs, tussocks, and Polylepis trees. We chose areas with slopes between 20% and 50%, avoiding sites with steep topography, based on the slope map generated from the Digital Elevation Model (NASA SRTM Digital Elevation 30m). In these areas, we established 30 circular plots of 0.07 ha (707 m2) following the methodology of the National Forest Inventory (NFI) for forests in Colombia (Barreto et al., \u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e2018\u003c/span\u003e), adapted to analyze the heterogeneity of p\u0026aacute;ramo vegetation. The adaptation involved placing a 2x2 m subplot at the center of each circular plot for the measurement of non-woody plants and caulescent rosette species. We followed the growth form classification for p\u0026aacute;ramo vegetation: shrubs, basal rosettes, tussocks, herbs and caulirosettes (Cleef \u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e1981\u003c/span\u003e; Hughes and Atchison \u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e2015\u003c/span\u003e; Ramsay and Oxley \u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e2001\u003c/span\u003e). We measured small shrubs and trees (2.5 cm\u0026thinsp;\u0026le;\u0026thinsp;DBH\u0026thinsp;\u0026lt;\u0026thinsp;10 cm) in plots within 3 m radius; medium trees (10 cm\u0026thinsp;\u0026le;\u0026thinsp;DBH\u0026thinsp;\u0026lt;\u0026thinsp;30 cm) in subplots within a 7 m radius, and large trees (DBH\u0026thinsp;\u0026ge;\u0026thinsp;30 cm) within a 15 m radius (Barreto et al. \u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e2018\u003c/span\u003e).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eWe measure plant height (H, in meters) of all recorded individuals using a Vertex IV Hagl\u0026ouml;f and measuring tapes (0.1 cm precision). The diameter of each individual was measured using a caliper (0.01 cm precision) and measuring tapes. The circumference was converted to diameter using the formula GBH/π. In the 2x2 m subplots, we measured individuals with growth forms such as herbs, cushions, basal rosettes, caulescent rosettes, and tussock grasses, recording their height, diameter, and foliage area. we measured the diameter at 0.5 cm from the ground, the total height from the ground surface to the apical leaf, the stem height was measured from the ground surface to the base where the rosette begins, and the leaf area. For the caulirosettes, cushions, basal rosettes, tussock grasses, and herbs, the leaf area was measured in the North-South and East-West directions from the leaf tips. To avoid altering the condition of the plots and biomass contents for long-term monitoring, we collected individuals outside the plot at a maximum distance of 5 meters. Finally, we transported the samples to the Ecosystems and Climate Change Laboratory at Universidad Javeriana in Bogot\u0026aacute;, where they were stored in individually labeled paper bags and placed in an oven at 60\u0026deg;C for 48 hours until their dry weight stabilized, which was recorded for each individual. Necromass was not included, as it consists of decaying tissue attached to the plant; we included leaves and stems and removed reproductive structures (K\u0026ouml;rner \u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e2021\u003c/span\u003e). We collected all vegetation classified as herbs and we collected individuals for all other growth forms. We placed these plants in bags and dried in an oven at 60\u0026deg;C for 48 hours until reaching a constant weight and were subsequently weighed on a precision balance in the Ecosystems and Climate Change laboratory at Pontificia Universidad Javeriana.\u003c/p\u003e \u003cp\u003eAllometric models\u003c/p\u003e \u003cp\u003eFor each growth form, we collected 167 shrubs, 56 cushion plants, 89 stem rosettes, 32 tussock grasses, 63 basal rosettes, and 54 herbs. These individuals were collected in the El Cocuy National Natural Park and other protected p\u0026aacute;ramo areas of the eastern and central Colombian mountain ranges between 2017 and 2024 (Chingaza National Natural Park and Los Nevados National Natural Park). For each growth form, we established different allometric models considering the relationship between biomass and the predictive variables: total height, diameter, stem area, basal cylindrical volume (VCB\u0026thinsp;=\u0026thinsp;π* radius* height) and elliptical cone volume (VCO\u0026thinsp;=\u0026thinsp;1/3* π *radius2*height) for tussocks and caulirosettes (Johnson et al. \u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e1988\u003c/span\u003e; McClaran et al. \u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e2013\u003c/span\u003e) (see appendix 2). We tested these equations using simple power-law functions (Y\u0026thinsp;=\u0026thinsp;a\u0026thinsp;+\u0026thinsp;xb) for each predictor, including the relationship between them. We tested the distribution of each variable to ensure the assumption of normality and homoscedasticity. Variables that deviate from normality were transformed using the logarithm or square root. Model selection for each growth form was based on a combination of visual inspection of residuals, outliers, and influential points for each model, as well as statistical metrics such as AIC weights, mean squared error, and adjusted R\u0026sup2; (Zuur et al. \u003cspan citationid=\"CR60\" class=\"CitationRef\"\u003e2009\u003c/span\u003e). We calculated the mean squared error using k-fold cross-validation (Bro et al. \u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e2008\u003c/span\u003e); in which 60% of the data was selected for model calibration, and the remaining 40% was used for error estimation. These values are reported for each allometric model and growth form.\u003c/p\u003e \u003cp\u003ePlot biomass estimation\u003c/p\u003e \u003cp\u003eAccording to the best models selected for each growth form, the equations were applied to the individuals recorded in each plot, except for large tree including Polylepis where used already published equations were used (Alvarez et al. \u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2012\u003c/span\u003e; V\u0026aacute;squez et al. \u003cspan citationid=\"CR54\" class=\"CitationRef\"\u003e2014\u003c/span\u003e). Biomass estimates for each individual were back transformed to weight units using the relevant function exponential for those variables log transformed. An additional correction was performed using the correction factor (CF), based on the standard residual sum of squares (RSE\u0026sup2;) for each selected model (Baskerville \u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e1972\u003c/span\u003e; Chave et al. \u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e2005\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eEach plot was analyzed using a Light Detection and Ranging (LiDAR) sensor, specifically the Zenmuse L1 mounted on a DJI Matrice 300 RTK UAV. We selected detailed parameters for the flight mission, including flight pattern, area, altitude, trajectory, speed, and point density, accordingly the conditions of the study sites (Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e). Due to the heterogeneity of the study areas, flight altitudes ranged from 50 to 70 meters, with higher altitudes used in areas where vegetation was taller. Additionally, we conducted flights at lower speeds to obtain dense point clouds, which allowed for the identification of small vegetation such as grasses and herbs (Zhang et al. \u003cspan citationid=\"CR58\" class=\"CitationRef\"\u003e2021\u003c/span\u003e). To ensure comprehensive data coverage, we defined flight areas using polygons that encompassed all plots on the ground for each cover vegetation type, obtaining six flight zones, ranging from 3 to 10 hectares, based on the spatial distribution of ground plots. We conducted an oblique flight mode, capturing data from different angles to obtain detailed three-dimensional point clouds. We georeferenced each plot by locating its centroid using a high-precision RTK antenna and triangulating coordinates obtained from GPS. According to each flight zone, we applied a zonal statistic to the resulting biomass raster at the landscape scale.\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\u003eParameters of flight mission for the different vegetation covers in the Lagunillas paramo area using the Zenmuse L1 sensor on the Matrice 300 RTK UAV.\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\u003eFlight parameter\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eValue\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eFlight parameter\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eValue\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eFlight route mission\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eOblique\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e\u003cb\u003eFlight zone\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e3\u0026ndash;9 hectares\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eFlight speed\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e5m/seg\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e\u003cb\u003eFlight duration\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e15\u0026ndash;27 min\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eAverage LiDAR Density\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e950\u0026ndash;3.365 points/m\u003csup\u003e2\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e\u003cb\u003eGSD (Ground Sample Distance)\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e1\u0026ndash;3 cm/pixel\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eSensor size\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e152 x 110 x 169 mm\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e\u003cb\u003eBands\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eNear Infrared\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eAltitude\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e50\u0026ndash;70 meters\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e\u003cb\u003eDatum\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eWKID 9377\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003eLiDAR data processing\u003c/h2\u003e \u003cp\u003eWe processed LiDAR data using DJI Terra software to reconstruct the raw data into point clouds in .LAS format. We processed point clouds in RStudio software, using integrated algorithms from the LidR and ForestGARP packages (Roussel et al. \u003cspan citationid=\"CR45\" class=\"CitationRef\"\u003e2020\u003c/span\u003e). To generate the Digital Terrain Model (DTM), we classified the ground points through the Triangular Irregular Network (TIN), which uses Delaunay triangulation as a linear interpolation method to produce a more detailed DTM (Roussel et al. \u003cspan citationid=\"CR45\" class=\"CitationRef\"\u003e2020\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eWith the obtained DTM, we normalized the point cloud to classify the data corresponding to vegetation. Subsequently, we generated the Canopy Height Model (CHM) using the Point to Raster (P2R) algorithm. This algorithm assigns the height of the highest point within its area to each pixel in the resulting raster, applying an adjustment that considers 8 surrounding points to the original. This forms a sub circle to treat each LiDAR point as a disk rather than an individual point. This process densifies the point cloud and produces a more uniform CHM with fewer empty pixels (Roussel et al. \u003cspan citationid=\"CR45\" class=\"CitationRef\"\u003e2020\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eUsing the CHM and applying the Individual Tree Detection (ITD) algorithm from the LidR package (Roussel et al. \u003cspan citationid=\"CR45\" class=\"CitationRef\"\u003e2020\u003c/span\u003e), we identified individual vegetation by segmenting the canopy. We obtained structural metrics such as maximum canopy height (Zmax), mean canopy height (Zmean), canopy height distribution by percentiles (Zp75), as well as high (Imax), mean (Imean), and percentile-distributed intensity for each segmented individual (Ipcmz10, Ipcmz50, and Ipcmz90).\u003c/p\u003e \u003cp\u003eUAV LiDAR metrics\u003c/p\u003e \u003cp\u003eWe estimated Lorey's Height as the index that calculates the average height of each plant individual in proportion to its basal area for each plot in each cover type. This is a relevant parameter derived from LiDAR data, particularly for ecosystems with heterogeneous structure (Rajab Pourrahmati et al. \u003cspan citationid=\"CR42\" class=\"CitationRef\"\u003e2018\u003c/span\u003e). We obtained the parameters for maximum Lorey height (Zmax), mean Lorey height (Zmean), and the 75th percentile height (Z75) for each plant individual by relating the convex area of the canopy to the heights (Roussel et al. \u003cspan citationid=\"CR45\" class=\"CitationRef\"\u003e2020\u003c/span\u003e). Additionally, we estimated the average height per plot using the Canopy Height Model (CHM). A filter was applied to the segmented vegetation individuals to exclude those composed of fewer than 100 points, with heights below 0.5 m or above 15 m. However, individuals with a higher number of points were retained to represent vegetation shorter than 0.5 m.\u003c/p\u003e \u003cp\u003eFor each plot within the different cover types, we obtained the Vegetation Cover fraction (VC) (Zhang et al., \u003cspan citationid=\"CR58\" class=\"CitationRef\"\u003e2021\u003c/span\u003e) to derive various metrics for shorter growth forms such as herbs, grasses, basal rosettes, cushion plants, and some caulescent rosettes. We obtained this metric from the CHM and segmented vegetation using the ITD algorithm, producing a factor ranging from 0 to 1. We measured the return intensity, which represents the amount of energy reflected back to the sensor as a function of the illuminated area and canopy reflectance (Garc\u0026iacute;a et al. \u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e2010\u003c/span\u003e), according to the equation proposed by Roussel et al. (2024) we obtained the intensity for each segmented plant individual and expressed as the 10th percentile (Ipcmz10), the 50th percentile (Ipcmz50), the 90th percentile (Ipcmz90) and its standard deviation (IpcmzSD). To examine the relationship between metrics obtained from LiDAR and data measured in the field, we performed a Pearson correlation analysis.\u003c/p\u003e \u003cp\u003eUAV LiDAR above-ground biomass\u003c/p\u003e \u003cp\u003eWe estimated above-ground biomass at the plot scale by developing linear regression models, using the above-ground biomass values derived from allometric regressions in each plot as the observed variable and the LiDAR metrics as explanatory variables, while analyzing their collinearity. We tested these equations using simple power law functions (Y\u0026thinsp;=\u0026thinsp;a\u0026thinsp;+\u0026thinsp;xb). We selected the model based on residual standard error, adjusted R\u0026sup2; (R\u0026sup2; adj), AIC, and visual inspection of the residual plots. These values are reported for each model (see Appendix 2).\u003c/p\u003e \u003cp\u003eWe evaluated the error in above-ground biomass estimation using the LiDAR sensor by calculating the Root Mean Square Error (RMSE), which provides a quantitative assessment of the performance of the LiDAR models in predicting vegetation biomass (Bazzo et al. \u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). A lower RMSE value indicates that the model's predictions are closer to the observed values, suggesting higher accuracy in the estimates and, therefore, better predictive quality in the regression model.\u003c/p\u003e \u003cp\u003eUsing the equation obtained from the model and the metrics derived from the sensor data processing (CHM_mean), we calculated above-ground biomass at landscape scale, generating a raster with 0.1m pixels. We performed zonal statistics for the total area covered by the UAV LiDAR for each flight zone, averaging the pixel values from the above-ground biomass raster in units of MgCha⁻\u0026sup1;. Pixels corresponding to surface water bodies, such as lakes and lagoons, were assigned NA values, meaning that only vegetation and bare soil were included in the zonal statistics.\u003c/p\u003e \u003c/div\u003e"},{"header":"Results","content":"\u003cp\u003eAllometric models\u003c/p\u003e \u003cp\u003eThe best models based on AIC, R\u0026sup2;, and the comparison of observed vs. predicted values for the growth forms of caulescent rosettes (R\u0026sup2;= 0.52), tussocks (R\u0026sup2;= 0.52), and cushions (R\u0026sup2;= 0.63) were those using leaf area as a predictor. For herbs (R\u0026sup2;= 0.43) and rosettes (R\u0026sup2;= 0.76), the best allometric models were those related the product of leaf area and plant height, while for shrubs, the best model contains the product of height and diameter as predictors (R\u0026sup2;= 0.60) (Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e). The errors associated with cross-validation for most growth forms were below one gram.\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\u003eSelected allometric models for estimating aboveground biomass (AGB) for each growth form of paramo vegetation. Predictors: H (total plant height in cm); FA (leaf area); D (diameter in cm), N (number of plants of the allometric set). Model coefficients: a, b, c, d. AIC: Akaike Information Criterion, R\u0026sup2;adj: adjusted R-squared, CV: Error obtained from cross-validation. The average and range of plant height and diameter variables used for calibrating the equations for each growth form in Paramo areas of the Colombian Andes.\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"12\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c7\" colnum=\"7\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c8\" colnum=\"8\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c9\" colnum=\"9\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c10\" colnum=\"10\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\"\u0026plusmn;\" class=\"colspec\" colname=\"c11\" colnum=\"11\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c12\" colnum=\"12\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGrowth form\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eRegression model\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eN\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003ea\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003eb\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c6\"\u003e \u003cp\u003ec\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c7\"\u003e \u003cp\u003ed\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c8\"\u003e \u003cp\u003eAIC\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c9\"\u003e \u003cp\u003eR2-adj\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c10\"\u003e \u003cp\u003eCV Error\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c11\"\u003e \u003cp\u003ePlant Height (McClaran et al.)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c12\"\u003e \u003cp\u003eDiameter\u003c/p\u003e \u003cp\u003e(McClaran et al.)\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eCaulirosette\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eLn(AGB)\u0026thinsp;=\u0026thinsp;a\u0026thinsp;+\u0026thinsp;b*sqrt(AF)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e46\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e3.304\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.075\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e193\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e0.52\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c10\"\u003e \u003cp\u003e1.1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\"\u0026plusmn;\" colname=\"c11\"\u003e \u003cp\u003e86.3\u0026thinsp;\u0026plusmn;\u0026thinsp;67 (9.6-259.8)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c12\"\u003e \u003cp\u003e29.23 (5.3\u0026ndash;76)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eTussocks\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eLn(AGB)\u0026thinsp;=\u0026thinsp;a\u0026thinsp;+\u0026thinsp;b*sqrt(AF)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e19\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e3.011\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.028\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e28.63\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e0.52\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c10\"\u003e \u003cp\u003e0.56\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\"\u0026plusmn;\" colname=\"c11\"\u003e \u003cp\u003e63\u0026thinsp;\u0026plusmn;\u0026thinsp;5 (33\u0026ndash;79)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c12\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eShrubs\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eLn (AGB)\u0026thinsp;=\u0026thinsp;a\u0026thinsp;+\u0026thinsp;b*log(H)\u0026thinsp;+\u0026thinsp;c*log(D)\u0026thinsp;+\u0026thinsp;d*log(H)*log(D)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e102\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e-0.290\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.810\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e-0.892\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.269\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e203.5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e0.60\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c10\"\u003e \u003cp\u003e0.58\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\"\u0026plusmn;\" colname=\"c11\"\u003e \u003cp\u003e64\u0026thinsp;\u0026plusmn;\u0026thinsp;4 (18\u0026ndash;250\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c12\"\u003e \u003cp\u003e8.5\u0026thinsp;\u0026plusmn;\u0026thinsp;18 (0.1-11.42\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eCushions\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eLn(AGB)\u0026thinsp;=\u0026thinsp;a\u0026thinsp;+\u0026thinsp;b*log(AF)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e32\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.59\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.403\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e33,40\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e0.63\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c10\"\u003e \u003cp\u003e0.25\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\"\u0026plusmn;\" colname=\"c11\"\u003e \u003cp\u003e7\u0026thinsp;\u0026plusmn;\u0026thinsp;10 (1-44.2)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c12\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eRosettes\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eLn(AGB)\u0026thinsp;=\u0026thinsp;a\u0026thinsp;+\u0026thinsp;b*sqrt(H)*c*log(AF)\u0026thinsp;+\u0026thinsp;d*sqrt(H)*log(AF)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e41\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e-0.770\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.628\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.113\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e-0.018\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e64.8\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e0.76\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c10\"\u003e \u003cp\u003e0.68\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\"\u0026plusmn;\" colname=\"c11\"\u003e \u003cp\u003e60\u0026thinsp;\u0026plusmn;\u0026thinsp;55 (1-104)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c12\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eHerbs\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eLn(AGB)\u0026thinsp;=\u0026thinsp;a\u0026thinsp;+\u0026thinsp;b*sqrt(H)\u0026thinsp;+\u0026thinsp;c*log(AF)\u0026thinsp;+\u0026thinsp;d*sqrt(H)*log(AF)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e25\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e-1.819\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.283\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e1.086\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e-0.09\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e81.23\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e0.43\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c10\"\u003e \u003cp\u003e0.55\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\"\u0026plusmn;\" colname=\"c11\"\u003e \u003cp\u003e40\u0026thinsp;\u0026plusmn;\u0026thinsp;14.5 (18\u0026ndash;87)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c12\"\u003e \u003cp\u003e0.5\u0026thinsp;\u0026plusmn;\u0026thinsp;0.12 (0.3\u0026ndash;0.6)\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\u003eUAV LiDAR metrics\u003c/p\u003e \u003cp\u003eBased on the individual tree segmentation within the plots, a minimum of 50 individuals were detected for tree and shrub cover, and up to 170 individuals in vegetation dominated by grasses. The relationship between field-measured heights and heights derived from Lorey\u0026rsquo;s height and the canopy height model (CHM) were very similar (R\u0026sup2;= 0.72, 0.73, respectively). With the vegetation cover metric, a greater dispersion of samples and a weaker relationship were observed (R\u0026sup2;= 0.24), while the intensity at the 10th percentile showed negative correlations (R\u0026sup2;= -0.43), which tended to decrease as the percentile increased. The correlation between aboveground biomass (AGB) obtained through allometric models and height metrics, such as Lorey\u0026rsquo;s height and the CHM, showed a similar trend to that observed with field-measured heights (R\u0026sup2;= 0.65 and R\u0026sup2;= 0.66, respectively) (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e) (Appendix 3).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eEstimation of aboveground biomass using allometry and UAV LiDAR\u003c/p\u003e \u003cp\u003eThe best model for estimating aboveground biomass at the landscape level using LiDAR is based on the mean canopy height model (mean CHM) and an error around two tons of carbon per hectare (R\u0026sup2; = 0.65- RMSE\u0026thinsp;=\u0026thinsp;2.29). This equation allows calculating the biomass for the different type of covers found in the p\u0026aacute;ramo zones, regardless of their structure (Table\u0026nbsp;\u003cspan refid=\"Tab3\" class=\"InternalRef\"\u003e4\u003c/span\u003e). The biomass expressed through the selected cover types represents the entire heterogeneity of the p\u0026aacute;ramo landscape, where caulirosettes, tussocks, shrub, and trees dominate different areas. Based on the allometric equations, vegetation dominated by shrubs, stored an aboveground biomass of 4.3\u0026thinsp;\u0026plusmn;\u0026thinsp;2.9 Mg C ha⁻\u0026sup1;. For areas dominated by stem rosettes, contained 3.8\u0026thinsp;\u0026plusmn;\u0026thinsp;2.0 Mg ha⁻\u0026sup1; in aboveground biomass. Areas dominates by tussocks and grasses retained 4.3\u0026thinsp;\u0026plusmn;\u0026thinsp;2 Mg C ha⁻\u0026sup1; of biomass, while areas dominated by Polylepis trees accumulated around 11.76\u0026thinsp;\u0026plusmn;\u0026thinsp;7 Mg C ha⁻\u0026sup1; of biomass.\u003c/p\u003e \u003cp\u003eUsing LiDAR metrics, vegetation dominated by shrub accumulated on aboveground biomass 5.4\u0026thinsp;\u0026plusmn;\u0026thinsp;1.3 Mg C ha⁻\u0026sup1;, caulirosette-dominated cover accumulated 3.7\u0026thinsp;\u0026plusmn;\u0026thinsp;0.5 Mg C ha⁻\u0026sup1;, a value very similar to that reported by the allometric equations. The aboveground biomass accumulated in areas dominated by tussocks, was 3.3\u0026thinsp;\u0026plusmn;\u0026thinsp;0.4 Mg C ha⁻\u0026sup1;. Finally, areas dominated by Polylepis stored a biomass of 12.1\u0026thinsp;\u0026plusmn;\u0026thinsp;4.2 Mg C ha⁻\u0026sup1;. These results show a pattern similar to those reported by the allometric equations (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003e).\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab3\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 4\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eModel estimators for the estimation of aboveground biomass at the landscape level in the p\u0026aacute;ramo, considering the CHM mean variable (average vegetation cover) obtained with LiDAR. AGB: aboveground biomass, Model coefficients: a, b, c, d. AIC: Akaike Information Criterion, R\u0026sup2;adj: adjusted R-squared, RMSE: root mean square error of the model.\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"7\"\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 \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 \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cem\u003eCobertura\u003c/em\u003e\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003e\u003cem\u003eRegression model\u003c/em\u003e\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003e\u003cem\u003ea\u003c/em\u003e\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003e\u003cem\u003eb\u003c/em\u003e\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003e\u003cem\u003eAIC\u003c/em\u003e\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c6\"\u003e \u003cp\u003e\u003cem\u003eR2-adju\u003c/em\u003e\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c7\"\u003e \u003cp\u003e\u003cem\u003eRMSE\u003c/em\u003e\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eGeneral\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eAGB\u0026thinsp;=\u0026thinsp;a + (b*CHM_mean)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e2.56\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e3.26\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e140.9\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0.65\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e2.29\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eLandscape aboveground estimation reflected the vegetation structure of each cover type. In flight zone 1, the average AGB was the highest compared to other areas due to the dominance of Polylepis trees, which have greater height compared to other cover types. Flight zones 3 and 5, predominantly covered by shrubs, showed lower average AGB values compared to zone 1, but higher values when comparing with zones 2, 5 and 6, in which shorter shrubs, grasses, and caulirosettes dominated (Table\u0026nbsp;\u003cspan refid=\"Tab4\" class=\"InternalRef\"\u003e5\u003c/span\u003e). We could determine carbon storage distribution for each flight zone with a very detailed spatial resolution (0.10 m resolution raster) by estimating AGB at landscape scale (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003e).\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 5\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eAboveground biomass (AGB) estimated at the landscape scale for each cover type in de 6 flight zones based on metrics obtain from LiDAR. We present the flight area covered, the aboveground biomass obtained, and the aboveground biomass estimated at the landscape scale.\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\u003e\u003cem\u003eFlight\u003c/em\u003e\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003e\u003cem\u003eArea (ha)\u003c/em\u003e\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003e\u003cem\u003eAGB (Mg C ha\u003c/em\u003e\u003csup\u003e\u003cem\u003e\u0026minus;\u0026thinsp;1\u003c/em\u003e\u003c/sup\u003e\u003cem\u003e)\u003c/em\u003e\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003e\u003cem\u003eAGB land scale (Mg C)\u003c/em\u003e\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eFlight Zone1\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e\u003cem\u003e3.47\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e\u003cem\u003e8.20\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e\u003cem\u003e28.45\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eFlight Zone2\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e\u003cem\u003e3.43\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e\u003cem\u003e3.37\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e\u003cem\u003e11.55\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eFlight Zone3\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e\u003cem\u003e9.17\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e\u003cem\u003e4.45\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e\u003cem\u003e40.8\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eFlight Zone4\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e\u003cem\u003e2.95\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e\u003cem\u003e3.37\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e\u003cem\u003e9.94\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eFlight Zone5\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e\u003cem\u003e10.58\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e\u003cem\u003e3.84\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e\u003cem\u003e40.62\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eFlight Zone6\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e\u003cem\u003e5.97\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e\u003cem\u003e3.27\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e\u003cem\u003e19.52\u003c/em\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\u003e \u003c/p\u003e"},{"header":"Discussion","content":"\u003cp\u003eThis study is a pioneering in estimating aboveground biomass in tropical high mountain landscapes using allometric models and LiDAR-derived equations. The results suggest that these models have the potential to capture the heterogeneity of vegetation in p\u0026aacute;ramo landscapes, as they accurately predict the aboveground biomass across diverse vegetation structure. Our findings from allometric equations and LiDAR suggest that the tropical p\u0026aacute;ramos of Colombia can store between 3 y 5 Mg C ha\u003csup\u003e\u0026minus;\u0026thinsp;1\u003c/sup\u003e. These values are consistent with values reported in other alpine areas in the Andes with 3.3 Mg C ha\u003csup\u003e\u0026minus;\u0026thinsp;1\u003c/sup\u003e for the grasslands of the Peruvian puna; 4.2 Mg C ha\u003csup\u003e\u0026minus;\u0026thinsp;1\u003c/sup\u003e in Ecuadorian p\u0026aacute;ramos, and the 4 o 5 Mg C ha\u003csup\u003e\u0026minus;\u0026thinsp;1\u003c/sup\u003e in the p\u0026aacute;ramos of the Eastern Cordillera of Colombia (Hofstede et al. \u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e1995\u003c/span\u003e; Oliveras et al. \u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e2014a\u003c/span\u003e; Ramsay and Oxley \u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e2001\u003c/span\u003e). This pattern in aboveground biomass in the p\u0026aacute;ramo may be related to factors such as vegetation density and species life span, which directly influence decomposition rates and organic matter production in ecosystems (Pinos et al. \u003cspan citationid=\"CR41\" class=\"CitationRef\"\u003e2017\u003c/span\u003e). Additionally, factors such as changes in soil moisture and nutrients, along with species composition, also play a crucial role in these variations (Cabrera and Duivenvoorden \u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e2020\u003c/span\u003e; Gao et al. \u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e2021\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eThe observed higher values of above-ground biomass in areas dominated by shrub and Polylepis (between 5 and 12 Mg C ha⁻\u0026sup1;), might be associated with the increase in wood production, similar to that has been observed in the p\u0026aacute;ramos of the Eastern range of Colombia (Cardozo and Schnetter \u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e1976\u003c/span\u003e). These values fall within the range reported for shrublands in the Himalayas, an alpine vegetation type similar to the tropical p\u0026aacute;ramo, with 6 Mg C ha⁻\u0026sup1; (1447,31 g m⁻\u0026sup2;) aboveground biomass (Nie et al. \u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e2018\u003c/span\u003e). In contrast, our biomass estimates for \u003cem\u003ePolylepis\u003c/em\u003e were lower than those reported in Ecuador (14 and 200 Mg C ha⁻\u0026sup1;) (Montalvo et al. \u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e2018\u003c/span\u003e). This discrepancy may be attributed to differences in environmental conditions such as altitude, climate and soil characteristics. Moreover, we observed a lower density of Polylepis individuals per hectare than the ones observed in Ecuador (Montalvo et al. \u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e2018\u003c/span\u003e), with an average of 48 individuals/hectare compared to over 100 individuals/hectare respectively. Understanding the factors driving this variation might be a further step, which is crucial for accurately assessing carbon sequestration potential in the p\u0026aacute;ramo region.\u003c/p\u003e \u003cp\u003eTo improve the accuracy of biomass estimates, we developed multi-species allometric models for each growth form separately. These equations incorporated the variation in morphological characteristics and biomass accumulation, reflecting the different carbon assimilation mechanisms that determine each growth form (Dorrepaal \u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e2007\u003c/span\u003e). For most growth forms, allometric models relied on predictors such as height and leaf area (Tol and Cleef \u003cspan citationid=\"CR52\" class=\"CitationRef\"\u003e1994\u003c/span\u003e). Similarly, for tussocks, leaf area was primary predictor of biomass, consistent with findings from other Andean grasslands studies (Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e) (McClaran et al. \u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e2013\u003c/span\u003e; Rojo et al. \u003cspan citationid=\"CR44\" class=\"CitationRef\"\u003e2017\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eFor shrubs, the best models included height and stem diameter as predictors, similar to the allometric models used for tropical trees (Alvarez et al. \u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2012\u003c/span\u003e; Chave et al. \u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e2005\u003c/span\u003e) and shrubs, where the combination of height and diameter improves model estimates. In contrast with other studies that use crown diameter (Grigal and Ohmann \u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e1977\u003c/span\u003e; Smith and Brand \u003cspan citationid=\"CR50\" class=\"CitationRef\"\u003e1983\u003c/span\u003e; Torres et al. \u003cspan citationid=\"CR53\" class=\"CitationRef\"\u003e2012\u003c/span\u003e). This may be due to many of the collected species having few branches, such as some shrub species in the genus \u003cem\u003eMonticalia\u003c/em\u003e. For herbs, although the R\u003csup\u003e2\u003c/sup\u003e is lower than reported in other studies, the error is below 1 g. The low R\u003csup\u003e2\u003c/sup\u003e may be explained by the small sample size used to calibrate the allometric model, which can contribute to increase model uncertainty, as a reduced number of samples decreases the predictive power of the models. Therefore, it is recommended to use a larger number of individuals (80\u0026ndash;100) (Chave et al. \u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e2005\u003c/span\u003e; Sileshi \u003cspan citationid=\"CR48\" class=\"CitationRef\"\u003e2014\u003c/span\u003e). This is a challenge for some growth forms where densities are low. Nevertheless, providing information for estimating biomass in Andean vegetation is essential for medium- and long-term carbon monitoring. By using growth form-specific allometric models, we were able to capture the variation in biomass-physiognomy relationship and obtain more reliable estimates of carbon storage in p\u0026aacute;ramo ecosystems.\u003c/p\u003e \u003cp\u003eStructural variables of p\u0026aacute;ramo vegetation detected through LiDAR showed a correlation with parameters measured in ground-plots, confirming findings from previous research that highlight the usefulness of three-dimensional LiDAR models for detecting vegetation variability, including small sized species such as grasses and herbs (Wang et al. \u003cspan citationid=\"CR56\" class=\"CitationRef\"\u003e2017\u003c/span\u003e; Zhang et al. \u003cspan citationid=\"CR59\" class=\"CitationRef\"\u003e2017\u003c/span\u003e). In this study, a correlation of 0.73 and 0.72 was observed between field-measured height and height derived from the Canopy Height Model (CHM) and Lorey\u0026rsquo;s Height, underscoring the importance of height as a key structural variable for estimating aerial biomass (AGB) in grassland ecosystems (Bazzo et al. \u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). Additionally, our results were consistent with those reported by Zhang et al. (\u003cspan citationid=\"CR58\" class=\"CitationRef\"\u003e2021\u003c/span\u003e) where LiDAR metrics showed a strong correlation with biomass obtained through allometry, height, and FVC. The individual segmentation of smaller vegetation remains a challenge, as it depends on parameters established during data acquisition with the sensor, such as flight speed, flight shape, and altitude, as well as the performance of the algorithm used during data processing. To address this challenge, our study planned flights at lower speeds and altitudes, which allowed for the acquisition of dense point clouds that facilitated the identification of low vegetation, such as grasses, herbs, and cushions plants (Zhang et al. \u003cspan citationid=\"CR58\" class=\"CitationRef\"\u003e2021\u003c/span\u003e). Additionally, algorithm parameters were determined to improve individual segmentation (Roussel et al. \u003cspan citationid=\"CR45\" class=\"CitationRef\"\u003e2020\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eThe structural complexity of grasslands, along with height and species richness, has been identified as an influencing factor in some spectral properties of vegetation (Villoslada et al. \u003cspan citationid=\"CR55\" class=\"CitationRef\"\u003e2020\u003c/span\u003e). To tackle this, we produced several metrics from LiDAR data, focusing on the structural diversity and heterogeneity of p\u0026aacute;ramo vegetation. The selected model was based on predictors derived from the mean plot-level heights obtained from the Canopy Height Model (CHM), which have proven effective across a wide range of grasslands globally (K\u0026uuml;mmerer et al. \u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e2023\u003c/span\u003e; Wang et al. \u003cspan citationid=\"CR56\" class=\"CitationRef\"\u003e2017\u003c/span\u003e; Zhang et al. \u003cspan citationid=\"CR58\" class=\"CitationRef\"\u003e2021\u003c/span\u003e; Zhang et al. \u003cspan citationid=\"CR59\" class=\"CitationRef\"\u003e2017\u003c/span\u003e). This approach allowed for the integration of different growth forms present in all the studied coverages, successfully predicting aerial biomass for the heterogeneity of the p\u0026aacute;ramo at a landscape scale with a coefficient of determination (R\u0026sup2;= 0.67) and an error of 2 Mg C ha⁻\u0026sup1;. These results are comparable to those obtained by Chan et al. (\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e2021\u003c/span\u003e) in subtropical forests and exceed those reported by Zhang et al. (\u003cspan citationid=\"CR58\" class=\"CitationRef\"\u003e2021\u003c/span\u003e) in Mongolian grasslands.\u003c/p\u003e \u003cp\u003eAs reported by Gonz\u0026aacute;lez-Jaramillo et al. (\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e2018\u003c/span\u003e) regarding biomass prediction using LiDAR in grasses and caulirosette-dominated coverages, certain limitations were evident related, showing an underestimation of above-ground biomass. For this vegetation, the segmentation of the vegetation structure is not as accurate compared to the segmentation of trees and shrubs, where their structure is well-defined by three-dimensional models. Particularly for caulirosettes, the variables in the watershed segmentation algorithm had limitations in identifying individual\u0026rsquo;s due to their morphology (Peitzsch et al. \u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). This might be related to the structure of the trunk, which is composed of dead leaves retained in form of necromass and for some other individuals, it may be related to a shorter or not fully developed stems. Their detection may require alternative algorithms that incorporate variables adapted to the structure of this type of vegetation.\u003c/p\u003e \u003cp\u003eAlthough these results are of great importance for proposing innovative methodologies for monitoring the carbon cycle of high mountain ecosystems in the Andes, allowing the evaluation of their status and in some cases, the efficiency of implemented passive or active restoration measures, the comparison of biomass estimates in different studies still presents challenges due to the lack of uniformity in sampling methodologies (Oliveras et al. \u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e2014b\u003c/span\u003e). Therefore, further studies with standardized methodologies are recommended to improve comparability and understanding of these ecosystems. It is still necessary to continue developing new models that integrate metrics obtained from complementary UAV sensors with allometric models, mainly for these ecosystems that present such heterogeneous vegetation.\u003c/p\u003e \u003cp\u003eThe allometric and LiDAR based models presented here can be applied in remote tropical alpine ecosystems where often the topography of the Andes can be a limiting factor, and remote sensing strategies that circumvent the logistical constraints, the difficulties of accessing remote locations and allows precise ABG carbon estimates in remote landscapes such as the Andean p\u0026aacute;ramos.\u003c/p\u003e"},{"header":"Conclusions","content":"\u003cp\u003eOur results confirm that UVA can be used to estimate the biomass of high mountain grasslands and shrublands; however the method could be improved for rosettes, specifically for species of the genus \u003cem\u003eEspeletia\u003c/em\u003e. Additionally, the great potential to estimate biomass at a regional scale with a high degree of accuracy allows for a more accurate estimation of above ground biomass and carbon accumulation patterns on an endangered ecosystem. Mos important, the methods open the possibility of monitoring large areas of paramo areas at low cost and high precision including restored, burned or grazed areas, offering an alternative to the different conservation agencies currently engaged in the protection of the p\u0026aacute;ramo.\u003c/p\u003e"},{"header":"Declarations","content":"\u003ch2\u003eAuthor Contribution\u003c/h2\u003e\u003cp\u003ePV: Formal analysis, research, visualization; Writing: original draft preparation and review and editing; data collection. AR: Formal analysis; research; visualization; Writing: original draft preparation and review and editing; data collection. LS: Formal analysis; research; visualization; Writing: original draft preparation, review, and editing. MS: Formal analysis; research; Writing: review and editing.JCB: Conceptualization; funding acquisition; supervision; Writing: review and editing. LFP: Conceptualization; funding acquisition; edititing. FN: Conceptualization; funding acquisition; edititing.\u003c/p\u003e\u003ch2\u003eAcknowledgement\u003c/h2\u003e\u003cp\u003eWe would like to thank the Pontificia Universidad Javeriana of Bogot\u0026aacute; for the opportunity to research this high mountain topic and for the use of its facilities. To Ecopetrol S.A and the ICEPT (Instituto Colombiano del Petr\u0026oacute;leo y Las Energ\u0026iacute;as de la Transici\u0026oacute;n) for its funding and support of the research. To the National Natural Park El Cocuy and all the park rangers for their support in the field. To Marcos Correa, Edward Avil\u0026aacute;n, Laura B\u0026aacute;ez, Silvio Orteg\u0026oacute;n, Edna Rinc\u0026oacute;n, Angie G\u0026oacute;mez, Freddy Avellaneda, Astrid Cruz, and David Hern\u0026aacute;ndez for their support in the field and laboratory activities.\u003c/p\u003e\n\u003ch3\u003eData availability\u003c/h3\u003e\n\u003cp\u003eThe authors declare that the data supporting the findings of this study are available within the paper and its Supplementary Information files. Should any raw data files be needed in another format they are available from the corresponding author upon reasonable request.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n \u003cli\u003eAlmeida DRAd, Stark SC, Shao G et al (2019) Optimizing the remote detection of tropical rainforest structure with airborne lidar: Leaf area profile sensitivity to pulse density and spatial sampling. Remote Sensing 11(1):92\u003c/li\u003e\n \u003cli\u003eAlvarez E, Duque A, Saldarriaga J et al (2012) Tree above-ground biomass allometries for carbon stocks estimation in the natural forests of Colombia. Forest Ecology and management 267:297-308\u003c/li\u003e\n \u003cli\u003eAnderson KE, Glenn NF, Spaete LP et al (2018) Estimating vegetation biomass and cover across large plots in shrub and grass dominated drylands using terrestrial lidar and machine learning. Ecological Indicators 84:793-802\u003c/li\u003e\n \u003cli\u003eAsner GP, Mascaro J (2014) Mapping tropical forest carbon: Calibrating plot estimates to a simple LiDAR metric. Remote Sensing of Environment 140:614-624\u003c/li\u003e\n \u003cli\u003eBarreto J, Capachero C, Barbosa A, Phillips J, Pe\u0026ntilde;a M (2018) Manual de Campo Inventario Forestal Nacional de Colombia.\u0026nbsp;\u003c/li\u003e\n \u003cli\u003eBaskerville G (1972) Use of logarithmic regression in the estimation of plant biomass. Canadian Journal of Forest Research 2(1):49-53\u003c/li\u003e\n \u003cli\u003eBazzo COG, Kamali B, H\u0026uuml;tt C, Bareth G, Gaiser T (2023) A review of estimation methods for aboveground biomass in grasslands using UAV. Remote Sensing 15(3):639\u003c/li\u003e\n \u003cli\u003eBro R, Kjeldahl K, Smilde AK, Kiers H (2008) Cross-validation of component models: a critical look at current methods. Analytical and bioanalytical chemistry 390:1241-1251\u003c/li\u003e\n \u003cli\u003eCabrera M, Duivenvoorden JF (2020) Drivers of aboveground biomass of high mountain vegetation in the Andes. Acta Oecologica 102:103504\u003c/li\u003e\n \u003cli\u003eCamarretta N, Harrison PA, Bailey T et al (2020) Monitoring forest structure to guide adaptive management of forest restoration: a review of remote sensing approaches. New Forests 51(4):573-596\u003c/li\u003e\n \u003cli\u003eCardozo H, Schnetter M-L (1976) Estudios ecol\u0026oacute;gicos en el P\u0026aacute;ramo de Cruz Verde, Colombia. III. La biomasa de tres asociaciones vegetales y la productividad de Calamagrostis effusa (HBK) Steud. y Paepalanthus columbiensis Ruhl. en comparaci\u0026oacute;n con la concentraci\u0026oacute;n de clorofila. Caldasia:69-83\u003c/li\u003e\n \u003cli\u003eChan E, Fung T, Wong F (2021) Estimating above-ground biomass of subtropical forest using airborne LiDAR in Hong Kong, Sci. Rep.-UK, 11, 1751.\u0026nbsp;\u003c/li\u003e\n \u003cli\u003eChave J, Andalo C, Brown S et al (2005) Tree allometry and improved estimation of carbon stocks and balance in tropical forests. Oecologia 145:87-99\u003c/li\u003e\n \u003cli\u003eChave J, Condit R, Aguilar S, Hernandez A, Lao S, Perez R (2004) Error propagation and scaling for tropical forest biomass estimates. Philosophical Transactions of the Royal Society of London. Series B: Biological Sciences 359(1443):409-420\u003c/li\u003e\n \u003cli\u003eCleef A (1981) The vegetation of the p\u0026aacute;ramos of the Colombian Cordillera Oriental. Mededelingen van het Botanisch Museum en Herbarium van de Rijksuniversiteit te Utrecht 481(1):1-320\u003c/li\u003e\n \u003cli\u003eDorrepaal E (2007) Are plant growth‐form‐based classifications useful in predicting northern ecosystem carbon cycling feedbacks to climate change? Journal of Ecology 95(6):1167-1180\u003c/li\u003e\n \u003cli\u003eFarley KA, Anderson WG, Bremer LL, Harden CP (2011) Compensation for ecosystem services: an evaluation of efforts to achieve conservation and development in Ecuadorian p\u0026aacute;ramo grasslands. Environmental Conservation 38(4):393-405\u003c/li\u003e\n \u003cli\u003eGao X, Dong S, Xu Y et al (2021) Plant biomass allocation and driving factors of grassland revegetation in a Qinghai‐Tibetan Plateau chronosequence. Land Degradation \u0026amp; Development 32(4):1732-1741\u003c/li\u003e\n \u003cli\u003eGarc\u0026iacute;a M, Ria\u0026ntilde;o D, Chuvieco E, Danson FM (2010) Estimating biomass carbon stocks for a Mediterranean forest in central Spain using LiDAR height and intensity data. Remote Sensing of Environment 114(4):816-830\u003c/li\u003e\n \u003cli\u003eGong K, Huang Z, Qu M et al (2024) Influences of climate change on carbon and water fluxes of the ecosystem in the Qinling Mountains of China. Ecological Indicators 166:112504\u003c/li\u003e\n \u003cli\u003eGonz\u0026aacute;lez-Jaramillo V, Fries A, Zeilinger J, Homeier J, Paladines-Benitez J, Bendix J (2018) Estimation of above ground biomass in a tropical mountain forest in southern Ecuador using airborne LiDAR data. Remote Sensing 10(5):660\u003c/li\u003e\n \u003cli\u003eGrigal D, Ohmann L (1977) Biomass estimation for some shrubs from northeastern Minnesota. North Central Forest Experiment Station, Forest Service. US Department of Agriculture, Research Note, NC-226 3\u003c/li\u003e\n \u003cli\u003eHeimann M, Reichstein M (2008) Terrestrial ecosystem carbon dynamics and climate feedbacks. Nature 451(7176):289-292\u003c/li\u003e\n \u003cli\u003eHofstede RG, Castillo MXM, Osorio CMR (1995) Biomass of grazed, burned, and undisturbed p\u0026aacute;ramo grasslands, Colombia. I. Aboveground vegetation. Arctic and Alpine Research 27(1):1-12\u003c/li\u003e\n \u003cli\u003eHughes CE, Atchison GW (2015) The ubiquity of alpine plant radiations: from the Andes to the Hengduan Mountains. New Phytologist 207(2):275-282\u003c/li\u003e\n \u003cli\u003eJohnson PS, Johnson CL, West NE (1988) Estimation of phytomass for ungrazed crested wheatgrass plants using allometric equations. Rangeland Ecology \u0026amp; Management/Journal of Range Management Archives 41(5):421-425\u003c/li\u003e\n \u003cli\u003eKerkhoff AJ, Enquist BJ (2009) Multiplicative by nature: why logarithmic transformation is necessary in allometry. Journal of Theoretical Biology 257(3):519-521\u003c/li\u003e\n \u003cli\u003eK\u0026ouml;rner C (2021) The alpine life zone. Alpine plant life: Functional plant ecology of high mountain ecosystems:23-51\u003c/li\u003e\n \u003cli\u003eK\u0026uuml;mmerer R, Noack PO, Bauer B (2023) Using high-resolution UAV imaging to measure canopy height of diverse cover crops and predict biomass. Remote Sensing 15(6):1520\u003c/li\u003e\n \u003cli\u003eLu D, Chen Q, Wang G, Liu L, Li G, Moran E (2016) A survey of remote sensing-based aboveground biomass estimation methods in forest ecosystems. International Journal of Digital Earth 9(1):63-105\u003c/li\u003e\n \u003cli\u003eMascaro J, Detto M, Asner GP, Muller-Landau HC (2011) Evaluating uncertainty in mapping forest carbon with airborne LiDAR. Remote Sensing of Environment 115(12):3770-3774\u003c/li\u003e\n \u003cli\u003eMcClaran M, McMurtry C, Archer S (2013) A tool for estimating impacts of woody encroachment in arid grasslands: allometric equations for biomass, carbon and nitrogen content in Prosopis velutina. Journal of Arid Environments 88:39-42\u003c/li\u003e\n \u003cli\u003eMinaya V, Corzo G, Romero-Saltos H et al (2016) Altitudinal analysis of carbon stocks in the Antisana p\u0026aacute;ramo, Ecuadorian Andes. Journal of Plant Ecology 9(5):553-563\u003c/li\u003e\n \u003cli\u003eMontalvo J, Minga D, Verdugo A et al (2018) Caracter\u0026iacute;sticas morfol\u0026oacute;gico-funcionales, diversidad arb\u0026oacute;rea, tasa de crecimiento y de secuestro de carbono en especies y ecosistemas de Polylepis del sur de Ecuador. Ecolog\u0026iacute;a austral 28(1):249-261\u003c/li\u003e\n \u003cli\u003eMu\u0026ntilde;oz Blanco FU, Ariano R, Buitrago VR, Reyes Leal M, Carvajal R (2005) Plan de Manejo Ambiental Parque Nacional Natural El Cocuy. Parques Nacionales Naturales de Colombia, Bogota, pp. 114\u003c/li\u003e\n \u003cli\u003eNie X-q, Yang L-c, Xiong F, Li C-b, Fan L, Zhou G-y (2018) Aboveground biomass of the alpine shrub ecosystems in Three-River Source Region of the Tibetan Plateau. Journal of Mountain Science 15(2):357-363\u003c/li\u003e\n \u003cli\u003eOliveras I, Eynden M, Malhi Y et al (2014a) Grass allometry and estimation of above‐ground biomass in tropical alpine tussock grasslands. Austral ecology 39(4):408-415\u003c/li\u003e\n \u003cli\u003eOliveras I, Girardin C, Doughty CE et al (2014b) Andean grasslands are as productive as tropical cloud forests. Environmental Research Letters 9(11):115011\u003c/li\u003e\n \u003cli\u003eOliveras I, van der Eynden M, Malhi Y et al (2014c) Grass allometry and estimation of above‐ground biomass in tropical alpine tussock grasslands. Austral ecology 39(4):408-415\u003c/li\u003e\n \u003cli\u003ePeitzsch EH, Martin-Mikle C, Hendrikx J, Birkeland K, Fagre D (2024) Characterizing vegetation and return periods in avalanche paths using lidar and aerial imagery. Arctic, Antarctic, and Alpine Research 56(1):2310333\u003c/li\u003e\n \u003cli\u003ePinos J, Studholme A, Carabajo A, Gracia C (2017) Leaf Litterfall and Decomposition of Polylepis reticulata in the Treeline of the Ecuadorian Andes. Mountain Research and Development 37(1):87-96\u003c/li\u003e\n \u003cli\u003eRajab Pourrahmati M, Baghdadi N, Darvishsefat AA et al (2018) Mapping Lorey\u0026rsquo;s height over Hyrcanian forests of Iran using synergy of ICESat/GLAS and optical images. European Journal of Remote Sensing 51(1):100-115\u003c/li\u003e\n \u003cli\u003eRamsay PM, Oxley ERB (2001) An assessment of aboveground net primary productivity in Andean grasslands of central Ecuador. Mountain Research and Development 21(2):161-167\u003c/li\u003e\n \u003cli\u003eRojo V, Arzamendia Y, P\u0026eacute;rez C, Baldo J, Vil\u0026aacute; B (2017) Double sampling methods in biomass estimates of Andean shrubs and tussocks. Rangeland Ecology \u0026amp; Management 70(6):718-722\u003c/li\u003e\n \u003cli\u003eRoussel J-R, Auty D, Coops NC et al (2020) lidR: An R package for analysis of Airborne Laser Scanning (ALS) data. Remote Sensing of Environment 251:112061\u003c/li\u003e\n \u003cli\u003eRuiz D, Moreno HA, Guti\u0026eacute;rrez ME, Zapata PA (2008) Changing climate and endangered high mountain ecosystems in Colombia. Science of the total environment 398(1-3):122-132\u003c/li\u003e\n \u003cli\u003eSalas EAL (2021) Waveform LiDAR concepts and applications for potential vegetation phenology monitoring and modeling: a comprehensive review. Geo-Spatial Information Science 24(2):179-200\u003c/li\u003e\n \u003cli\u003eSileshi GW (2014) A critical review of forest biomass estimation models, common mistakes and corrective measures. Forest Ecology and Management 329:237-254\u003c/li\u003e\n \u003cli\u003eSinde-Gonz\u0026aacute;lez I, Gil-Docampo M, Arza-Garc\u0026iacute;a M et al (2021) Biomass estimation of pasture plots with multitemporal UAV-based photogrammetric surveys. International Journal of Applied Earth Observation and Geoinformation 101:102355\u003c/li\u003e\n \u003cli\u003eSmith WB, Brand GJ (1983) Allometric biomass equations for 98 species of herbs, shrubs, and small trees. Research Note NC-299. St. Paul, MN: US Dept. of Agriculture, Forest Service, North Central Forest Experiment Station.\u0026nbsp;\u003c/li\u003e\n \u003cli\u003eTmu\u0026scaron;ić G, Manfreda S, Aasen H et al (2020) Current practices in UAS-based environmental monitoring. Remote Sensing 12(6):1001\u003c/li\u003e\n \u003cli\u003eTol GJ, Cleef AM (1994) Above-ground biomass structure of a Chusquea tessellata bamboo p\u0026aacute;ramo, Chingaza National Park, Cordillera Oriental, Colombia. Vegetatio 115:29-39\u003c/li\u003e\n \u003cli\u003eTorres AM, Pe\u0026ntilde;a EJ, Z\u0026uacute;\u0026ntilde;iga O, Pe\u0026ntilde;a JA (2012) Evaluaci\u0026oacute;n del impacto de actividades antr\u0026oacute;picas en el almacenamiento de carbono en biomasa vegetal en ecosistemas de alta monta\u0026ntilde;a de Colombia. Bolet\u0026iacute;n Cient\u0026iacute;fico. Centro de Museos. Museo de Historia Natural 16(1):132-142\u003c/li\u003e\n \u003cli\u003eV\u0026aacute;squez E, Ladd B, Borchard N (2014) Carbon storage in a high-altitude Polylepis woodland in the Peruvian Andes. Alpine botany 124:71-75\u003c/li\u003e\n \u003cli\u003eVilloslada M, Bergamo TF, Ward R et al (2020) Fine scale plant community assessment in coastal meadows using UAV based multispectral data. Ecological indicators 111:105979\u003c/li\u003e\n \u003cli\u003eWang D, Xin X, Shao Q, Brolly M, Zhu Z, Chen J (2017) Modeling aboveground biomass in Hulunber grassland ecosystem by using unmanned aerial vehicle discrete lidar. Sensors 17(1):180\u003c/li\u003e\n \u003cli\u003eYazaki T, Hirano T, Sano T (2016) Biomass accumulation and net primary production during the early stage of secondary succession after a severe forest disturbance in northern Japan. Forests 7(11):287\u003c/li\u003e\n \u003cli\u003eZhang X, Bao Y, Wang D et al (2021) Using uav lidar to extract vegetation parameters of inner mongolian grassland. Remote Sensing 13(4):656\u003c/li\u003e\n \u003cli\u003eZhang X, He G, Zhang Z, Peng Y, Long T (2017) Spectral-spatial multi-feature classification of remote sensing big data based on a random forest classifier for land cover mapping. Cluster Computing 20:2311-2321\u003c/li\u003e\n \u003cli\u003eZuur AF, Ieno EN, Walker NJ, Saveliev AA, Smith GM (2009) Mixed effects models and extensions in ecology with R. Springer\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":true,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"landscape-ecology","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"land","sideBox":"Learn more about [Landscape Ecology](https://www.springer.com/journal/10980)","snPcode":"10980","submissionUrl":"https://submission.nature.com/new-submission/10980/3","title":"Landscape Ecology","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"Springer Hybrid","inReviewEnabled":true,"inReviewRevisionsEnabled":false},"keywords":"above ground biomass, allometry, tropical alpine paramo, LiDAR, carbon dynamics","lastPublishedDoi":"10.21203/rs.3.rs-5348181/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-5348181/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003ch2\u003eContext\u003c/h2\u003e \u003cp\u003eP\u0026aacute;ramos, high-elevation alpine ecosystems found in the northern Andes, are a biodiversity hotspot and play a crucial role in climate change mitigation due to their carbon storage capacity. Above-ground biomass, AGB, serves as a key indicator of ecosystem health and carbon sequestration potential. Accurate estimates of above-ground biomass are essential for understanding the variability of carbon storage across different p\u0026aacute;ramo vegetation types, successional stages and degradation impacts supporting the design of effective conservation and management strategies.\u003c/p\u003e\u003ch2\u003eObjectives\u003c/h2\u003e \u003cp\u003eUsing a combination of methods from direct measurements to UAS LiDar, we describe the main patterns of above-ground biomass across contrasting vegetation types and plant growth forms in p\u0026aacute;ramos of the northeast Andes of Colombia.\u003c/p\u003e\u003ch2\u003eMethods\u003c/h2\u003e \u003cp\u003eThis study was conducted in conserved p\u0026aacute;ramo areas in El Cocuy National Natural Park in the northern part of the Colombian Andes. We measured the ABG biomass of the different growth forms and related that to relevant allometric traits by using simple linear models. Using the allometric equations we estimated the AGB of 30 plots in areas dominated by different p\u0026aacute;ramo vegetation types. Airborne LiDAR data was collected from these plots and canopy height and density metrics were processed to determine landscape-level above-ground biomass calibrated with the ground measurements.\u003c/p\u003e\u003ch2\u003eResults\u003c/h2\u003e \u003cp\u003eWe found that plant height, basal diameter, and leaf area explained above-ground biomass variation for the different growth forms. We selected models with canopy height model (CHM) as predictor, to explain above-ground biomass at the landscape lev el. Allometric and LiDAR derived models showed p\u0026aacute;ramo biomass values ranging from 3 to 11 Mg C ha\u003csup\u003e\u0026minus;\u0026thinsp;1\u003c/sup\u003e.\u003c/p\u003e\u003ch2\u003eConclusions\u003c/h2\u003e \u003cp\u003eOur results demonstrated that it is possible to understand above ground carbon accumulation patterns at the landscape level by combining direct and indirect methods, such as allometric equations and LiDAR data, in areas representing the heterogeneity of p\u0026aacute;ramo vegetation. This study is pioneering in providing information for non-forest carbon reservoirs and the impacts of human actions on the dynamics of the ABG biomass, which are crucial to reach national GHG emission targets.\u003c/p\u003e","manuscriptTitle":"Spatial Patterns of Above-Ground Biomass in Tropical Alpine Páramo Ecosystems Using Allometric Models and LiDAR Data","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2024-11-08 04:37:17","doi":"10.21203/rs.3.rs-5348181/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"Revision requested","date":"2025-01-26T07:11:11+00:00","index":"","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2025-01-15T17:11:51+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2024-12-24T03:26:40+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"36001672863246283380118448008838030302","date":"2024-12-18T04:39:32+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"228331658301461492333916692638694331125","date":"2024-11-07T12:22:45+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2024-11-05T22:16:39+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2024-11-05T22:11:55+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2024-10-29T09:59:39+00:00","index":"","fulltext":""},{"type":"submitted","content":"Landscape Ecology","date":"2024-10-28T15:03:56+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"landscape-ecology","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"land","sideBox":"Learn more about [Landscape Ecology](https://www.springer.com/journal/10980)","snPcode":"10980","submissionUrl":"https://submission.nature.com/new-submission/10980/3","title":"Landscape Ecology","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"Springer Hybrid","inReviewEnabled":true,"inReviewRevisionsEnabled":false}}],"origin":"","ownerIdentity":"a1485fbb-c90c-4954-8109-70daae3c7a9f","owner":[],"postedDate":"November 8th, 2024","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"published-in-journal","subjectAreas":[],"tags":[],"updatedAt":"2025-07-14T16:03:10+00:00","versionOfRecord":{"articleIdentity":"rs-5348181","link":"https://doi.org/10.1007/s10980-025-02159-0","journal":{"identity":"landscape-ecology","isVorOnly":false,"title":"Landscape Ecology"},"publishedOn":"2025-07-11 15:57:41","publishedOnDateReadable":"July 11th, 2025"},"versionCreatedAt":"2024-11-08 04:37:17","video":"","vorDoi":"10.1007/s10980-025-02159-0","vorDoiUrl":"https://doi.org/10.1007/s10980-025-02159-0","workflowStages":[]},"version":"v1","identity":"rs-5348181","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-5348181","identity":"rs-5348181","version":["v1"]},"buildId":"qtupq5eGEP_6zYnWcrvyt","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}