Effective identification of opportunities for agricultural Best Management Practices (BMPs) is essential for reducing nonpoint source pollution in intensively managed agricultural watersheds globally. In the Great Lakes basin, this task is complicated by widespread tile-drained agriculture. Existing decision-support tools often generate an overwhelming number of potential intervention sites without clear prioritization, which limits practical implementation. The Agricultural Conservation Planning Framework (ACPF) systematically maps potential BMP locations but provides limited guidance for prioritizing sites. To address this limitation, the present study developed and applied a framework that integrates ACPF with machine learning and explicit uncertainty mapping to predict a continuous, field-level measure of conservation need. A survey of 138 agricultural fields in a 205 km2 tile-drained watershed in Southern Ontario was conducted to derive a Composite BMP Need Score (CBNS), integrating field-observed evidence of BMP need, site severity, and land management indicators. Random Forest (RF) and Extreme Gradient Boosting (XGBoost) models were trained to predict CBNS using geospatial predictors. Both models exhibited comparable performance, with stable and consistent predictions across the independent test set (n = 35) and 100 repeated train-test splits (RMSE = 1.32 ± 0.12 CBNS units on a 0.5–6.5 scale). Learning curves indicated that performance was primarily limited by sample size rather than model capability. XGBoost was selected for spatial prediction due to its greater flexibility and enhanced ability to differentiate among fields, and was used to estimate CBNS for all 627 agricultural fields in the watershed. Model interpretation identified hydrologic connectivity and erosive force as the strongest drivers of conservation need. The framework identified the top 20% of agricultural fields requiring the most interventions. By combining machine learning with ACPF and incorporating uncertainty layers, this study advances precision conservation from conceptual understanding to prioritization guidance, offering a transparent and transferable framework applicable to agricultural watersheds worldwide, particularly in tile-drained landscapes.

The North American beaver (Castor canadensis) is expanding its distribution in the Arctic tundra. Due to the species' capacity to engineer ecosystems, they can transform surface water dynamics and biogeochemistry, permafrost stability, vegetation composition, and impact Indigenous subsistence practices. Understanding past temporal occupancy dynamics of beavers is essential for assessing beaver colonization dynamics and impacts on ecosystems and people, but in the absence of historical data, new methods are required to address such knowledge gaps. By felling vegetation, beavers leave a clear historical signal of their occupancy and by damming rivers, they can also create a hydrological signal of their colonization. We combined dendrochronological techniques with satellite image time-series analysis to infer beaver colonization and associated hydrological impacts at their northern range edge in the Inuvialuit Settlement Region in Arctic Canada. Over an similar to 130-km transect, we recorded 60 detected beaver lodge and/or dam sites. At three focal sites across a latitudinal gradient, we collected 94 Salix and Alnus shrub stems with browsing scars. To reconstruct browsing histories and infer past beaver occupancy, we cross-dated these against regional shrub-ring chronologies spanning 1973-2023 (Salix) and 1968-2023 (Alnus), formed from 99 unbrowsed shrubs collected across the survey transect. The browsing histories provide evidence of beaver colonization in the region starting in 2008. The northern site record indicates occupancy between 2008 and 2022, the central site between 2015 and 2022, and the southern site between 2011 and 2023. We compared findings at one lodge-dam complex site with a remote sensing-based time-series decomposition method for the detection and characterization of surface water changes. Abrupt change in surface water extent was detected directly upstream of dams between 2015 and 2019, aligning with beaver browsing activity between 2015 and 2021. Convergent results across two independent novel methods give strength to the approach used and provide a link between browsing activity and hydrological impacts. Our findings indicate that combined dendrochronology and satellite image time-series analysis may be applied elsewhere in the Arctic to gain insights into northward beaver range expansion and hydrological impacts in the understudied northern margins of their range.

Agricultural runoff is a major source of water pollution, particularly in tile-drained landscapes where fields are engineered to move water rapidly off the land, enabling earlier spring field operations. At present, few tools help farmers and watershed managers identify the most suitable best management practices (BMPs) to mitigate this nonpoint source pollution. For example, the Agricultural Conservation Planning Framework (ACPF) effectively identifies potential BMPs. However, it does not rank field-specific options or incorporate regional stakeholder priorities. To address this gap, this study developed a field-specific BMP assignment framework that converts conservation opportunity maps into ranked BMP recommendations by integrating hydrogeomorphic suitability, conservation-need classes, feasibility constraints, and stakeholder-adjustable policy weights. The framework was applied to 627 agricultural fields in the Medway Creek watershed, southern Ontario, using ACPF-derived spatial datasets, field-survey information from 138 fields, hydrogeomorphic clustering, a Composite BMP Need Score (CBNS), and multi-criteria BMP analysis. Four hydrogeomorphic field typologies were identified, and repeated K-means testing showed high cluster stability. Under a balanced multi-criteria setting, controlled drainage was the most frequent top-ranked practice, recommended for 292 fields, followed by grassed waterways, water and sediment control basins, wetlands, and riparian buffers. Sensitivity analysis and field-record consistency assessment supported the BMP ranking logic, indicating that stakeholder inputs refined the rankings without dominating the biophysically relevant suggestions. The field-specific thematic maps developed in this study provide spatially explicit decision support for prioritizing the right BMP in the right place, while final implementation requires local verification, engineering design, cost assessment, and landowner consultation.

In order to effectively reduce nonpoint source pollutants in agricultural areas within a watershed, a combination of Best Management Practices (BMPs) is selected based on their economic and environmental effectiveness. However, determining the optimal combination can be challenging due to the implementation costs and the consideration of decision makers' preferences. This research presents a methodology for integrating a genetic algorithm with the Annualized Agricultural Non-Point Source Pollution model (AnnAGNPS) to effectively select the most efficient BMPs placement for a given watershed. By optimizing BMPs placement, the model can minimize sediment loads from different types of erosion, including sheet/rill, ephemeral gully, and total erosion at the minimal cost. Results demonstrated that BMP placement by the optimization model reduced sediment load caused by sheet/rill by 84.6 %, ephemeral gully by 85.4 %, and total erosion by 86.3 % in the study watershed. Additionally, the model achieved these results at a minimal cost, making it a cost-effective solution for sediment load reduction in the watershed. Also, the results showed the effective implementation of the developed optimization approach for strategically locating BMPs in specific areas, rather than implementing them throughout the entire watershed. By targeting these areas and implementing suitable BMPs, the model was able to reduce the amount of sediment load and remain cost-effective. The proposed weighted overlay technique helped to place BMPs within agricultural fields instead of AnnAGNPS cells, making it easier for farmers to adopt and effectively reduce sediment load in each field. The developed model in the current study can be applied by decision makers in other watersheds with limited resources for implementing BMPs.

Information on lake Water Surface Elevation (WSE) in the Arctic permafrost region is essential for understanding the impacts of climate warming on water storage and hydrological processes. The location of these lakes, in remote areas with limited access to in-situ monitoring capacity, remains a significant constraint in retrieving WSE. To overcome these issues, we leveraged the high spatial along-track resolution of ICESat-2, the higher temporal resolution of Sentinel-3, and the monitoring accuracy of installed gauges for retrieving WSE time series for several Arctic lakes. We derived in-situ WSE benchmark from 20 out of the 36 HOBOware gauges installed on a series of Arctic lakes over the non-frozen period spanning mid-June to the end of September 2022. The ICESat-2 ATL13 product (2018-2024) and Sentinel-3 SRAL data sets (2016-2024) were acquired over the study area. A multiple linear regression model (MLRM) was used to estimate bias between the different data sources. The comparison between in-situ measured WSE and whole-lake ICESat-2 WSE estimates revealed a root-mean-square-difference (RMSD) of 0.36 m, although with positive bias from ICESat-2. Using our in-situ data and MLRM, we calibrated the ICESat-2 product for this Arctic Lake region that produced an RMSD of 0.03 m. With the goal of combining the two WSE products from the ICESat-2 and the Sentinel-3 satellites to attain a WSE time series we compared the satellite products. The RMSD observed between ICESat-2 and the Sentinel-3 was 1.7 m for this region. To improve the time series WSE, we calibrated the instruments to each other, resulting in an RMSD of 0.17 m between the two satellites' WSE. After producing a combined WSE from the calibrated ICESat-2 and Sentinel-3, an accuracy assessment with the DAHITI database WSE revealed that our product had an RMSD of 0.03 m. This study demonstrates the combination of these sources of data towards WSE monitoring in Arctic lakes experiencing permafrost-related changes. It is also expected to serve as a baseline inventory for the recently launched SWOT altimetry satellites in producing long-temporal monitoring for the sizes of lakes in Arctic permafrost regions.

Tile drainage systems are extensively implemented across the Laurentian Great Lakes Basin (GLB) to enhance agricultural productivity on poorly drained soils. However, these systems substantially contribute to excess nutrient runoff, particularly phosphorus (P) and nitrogen (N), exacerbating eutrophication and harmful algal blooms in the Great Lakes. This literature review synthesized current knowledge on nutrient loadings from tile-drained agricultural watersheds and evaluated the effectiveness of various agricultural best management practices (BMPs) in mitigating nutrient losses in the GLB. Through a meta-synthesis of field and watershed scale monitoring and modeling studies and statistical analysis using Box-Whisker plots and Monte Carlo simulations, we assessed the nutrient reduction potential of representative BMPs, including cover cropping, nutrient management, controlled drainage, and constructed wetlands in tile-drained landscapes. Findings indicated that individual BMPs substantially reduced nutrient loadings, but the effectiveness of these BMPs depended on site-specific factors, including climate conditions, soil type, and drainage system design. Integrated approaches at field, edge-of-field, and watershed scales with a combination of multiple BMPs enhanced nutrient reduction benefits, aligning with regional water quality targets. The review also highlighted the challenges of climate change that may undermine BMP performance by altering precipitation patterns and increasing extreme weather events. To address these complexities, we proposed a framework for developing adaptive BMP scenarios tailored to specific watershed conditions, emphasizing the need for long-term monitoring and hydrologic model enhancements. This framework was designed to help policymakers, stakeholders, and farmers protect water quality and balance agricultural productivity in the GLB and similar agricultural regions globally.

Repeatable methods capable of quantifying Arctic surface water extent at high resolutions are important, but still require development. Here, we present a study using very-high resolution (VHR) X-band Synthetic Aperture Radar (SAR) imagery from Capella Space for fine-scale semantic segmentation of Arctic surface water features. Our study proposes a modified U-Net encoder-decoder model for this task, optimized using the Nadam algorithm. Otsu thresholding was leveraged to rapidly generate 512 × 512-pixel patches for the U-Net, resulting in an efficient and automated training pipeline. Within this study, we also quantitatively compared the deep learning (DL) U-Net to a shallow machine learning (ML) algorithm, XGBoost (XGB), and evaluated the Capella Space imagery against spatially and temporally coincident Sentinel-1 C-band. Performance evaluations showed the U-Net outperforms XGB measured under several statistical metrics, reaching an Intersection over Union (IoU) of 0.955. An explainability analysis was conducted to complement this finding, using Gradient-weighted Class Activation Mapping (Grad-Cam). Visual analysis also underscored the extreme detail of small water features captured by Capella Space imagery, which are at times omitted or lack clarity in conventional Sentinel-1. This research makes several contributions to Arctic surface water mapping, demonstrating the effectiveness of combining VHR SAR imagery with DL.

Species' traits and environmental conditions determine the abundance of tree species across the globe. The extent to which traits of dominant and rare tree species differ remains untested across a broad environmental range, limiting our understanding of how species traits and the environment shape forest functional composition. We use a global dataset of tree composition of >22,000 forest plots and 11 traits of 1663 tree species to ask how locally dominant and rare species differ in their trait values, and how these differences are driven by climatic gradients in temperature and water availability in forest biomes across the globe. We find three consistent trait differences between locally dominant and rare species across all biomes; dominant species are taller, have softer wood and higher loading on the multivariate stem strategy axis (related to narrow tracheids and thick bark). The difference between traits of dominant and rare species is more strongly driven by temperature compared to water availability, as temperature might affect a larger number of traits. Therefore, climate change driven global temperature rise may have a strong effect on trait differences between dominant and rare tree species and may lead to changes in species abundances and therefore strong community reassembly.

Repeatable methods capable of quantifying Arctic surface water extent at high resolutions are important, but still require development. Here, we present a study using very-high resolution (VHR) X-band Synthetic Aperture Radar (SAR) imagery from Capella Space for fine-scale semantic segmentation of Arctic surface water features. Our study proposes a modified U-Net encoder-decoder model for this task, optimized using the Nadam algorithm. Otsu thresholding was leveraged to rapidly generate 512 x 512-pixel patches for the U-Net, resulting in an efficient and automated training pipeline. Within this study, we also quantitatively compared the deep learning (DL) U-Net to a shallow machine learning (ML) algorithm, XGBoost (XGB), and evaluated the Capella Space imagery against spatially and temporally coincident Sentinel-1 C-band. Performance evaluations showed the U-Net outperforms XGB measured under several statistical metrics, reaching an Intersection over Union (IoU) of 0.955. An explainability analysis was conducted to complement this finding, using Gradient-weighted Class Activation Mapping (Grad-Cam). Visual analysis also underscored the extreme detail of small water features captured by Capella Space imagery, which are at times omitted or lack clarity in conventional Sentinel-1. This research makes several contributions to Arctic surface water mapping, demonstrating the effectiveness of combining VHR SAR imagery with DL. Le d & eacute;veloppement des m & eacute;thodes reproductibles permettant de quantifier l'& eacute;tendue des eaux de surface de l'Arctique & agrave; fine r & eacute;solution est important. Nous pr & eacute;sentons ici une & eacute;tude utilisant des images radar & agrave; synth & egrave;se d'ouverture (SAR) en bande X & agrave; tr & egrave;s haute r & eacute;solution spatiale (THR) de Capella Space pour la segmentation s & eacute;mantique & agrave; & eacute;chelle fine des caract & eacute;ristiques des eaux de surface de l'Arctique. Notre & eacute;tude propose un mod & egrave;le codeur-d & eacute;codeur U-Net modifi & eacute; pour cette t & acirc;che, optimis & eacute; par l'algorithme Nadam. Le seuillage Otsu a & eacute;t & eacute; utilis & eacute; pour g & eacute;n & eacute;rer des imagettes de 512 x 512 pixels pour le mod & egrave;le U-Net, ce qui a permis de cr & eacute;er un pipeline d'apprentissage efficace et automatis & eacute;. Dans le cadre de cette & eacute;tude, nous avons & eacute;galement compar & eacute; quantitativement l'algorithme d'apprentissage profond U-Net (DL: Deep Learning) & agrave; un algorithme d'apprentissage automatique (ML: Machine Learning), XGBoost (XGB). De plus, nous avons compar & eacute; les r & eacute;sultats obtenus & agrave; partir de l'imagerie Capella Space avec ceux obtenus & agrave; partir des images Sentinel-1 en bande C, pour des r & eacute;solutions spatiales et temporelles similaires. Les & eacute;valuations ont montr & eacute; que les r & eacute;sultats issus du model U-Net surpassent ceux du mod & egrave;le XGB pour plusieurs m & eacute;triques statistiques, atteignant une intersection sur l'union (IoU: Intersection over Union) de 0,955. Une analyse plus approfondie a & eacute;t & eacute; r & eacute;alis & eacute;e pour mieux interpr & eacute;ter ces r & eacute;sultats, & agrave; l'aide de la cartographie d'activation de classe pond & eacute;r & eacute;e par gradient (Grad-Cam: Gradient-weighted Class Activation Mapping). L'analyse visuelle a & eacute;galement mis en & eacute;vidence comment l'imagerie Capella Space peut faire ressortir des petites caract & eacute;ristiques aquatiques, qui sont parfois omises ou manquent de clart & eacute; avec l'usage des images de Sentinel-1. Cette recherche apporte plusieurs contributions & agrave; la cartographie des eaux de surface de l'Arctique, d & eacute;montrant l'efficacit & eacute; de la combinaison de l'imagerie SAR THR avec un mod & egrave;le DL.

Hemlock woolly adelgid (HWA) is an invasive insect that affects the eastern hemlock population in North America, causing severe die-off and altering ecosystem dynamics. Understanding the distribution of eastern hemlock will improve future HWA management and protection of existing eastern hemlock populations. To determine the degree to which different forest types and species can be distinguished at the stand level with variable densities of eastern hemlock present, a Bayesian phenology model was used to compute seven phenological parameters from four spectral indices derived from Sentinel-2 time series imagery. We tested spectral and phenological parameters derived using this method across three classification levels, including broad forest type, hemlock density, and dominant or co-dominant evergreen species. Using Kruskal-Wallis with post-hoc Dunn’s test, we found that phenological parameters derived from the Inverted Red-Edge Chlorophyll Index and the Soil-Adjusted Vegetation Index provided the highest separability between groups across all three levels of classification. The seasonal minimum greenness and fall inflection day provided the highest degree of separability among hemlock density classes. Seasonal minimum greenness provided the highest degree of separability among evergreen species. Among the nine evergreen dominant or co-dominant species classes tested, hemlock stands were found to be separable from four of the classes. White pine stands and black spruce stands showed the highest degree of overall separability. This study demonstrates the potential for phenological parameters in stand-level evergreen species classification. The combination of Sentinel-2 time series and phenological modeling has the potential to enhance tree species mapping studies at regional scales.

In agricultural landscapes, the use of topographic index (TI) models has been common to predict the presence and extent of ephemeral gullies (EGs). However, these models face two significant challenges: (1) the accurate prediction of EGs relies heavily on a critical threshold (CT) value, which is difficult to determine optimally using existing strategies, and (2) the calibration of TI models limits their applicability on a larger scale. To address these limitations, the current study proposes two methods: (1) the division of the study area into zones based on key factors influencing gully formation, reducing the need for TI model calibration, and (2) a pixel-based binary classification approach coupled with a precision performance metric to identify the calibrated CT value within a watershed. The performance of seven TI models for predicting EG length was evaluated using local validation within zones and transferred validation between zones. Local validation demonstrated that among the TI models, modified stream power index (MSPI), stream power index (SPI), and compound topographic index (CTI), in descending order, yielded the most accurate predictions for EG length. Furthermore, the decrease in accuracy observed in the transferred MSPI model compared to the local MSPI model supported the study's hypothesis that dividing a large-scale area into distinct zones with varying topographic and climatic characteristics enables the determination of a CT value specific to each zone. Soil loss rates due to EGs ranged from 0.36 to 1 kg/m2 yr, aligning with findings from similar global studies. These findings offer valuable insight that can be integrated into comprehensive watershed and soil erosion models.

AimTo determine the relationships between the functional trait composition of forest communities and environmental gradients across scales and biomes and the role of species relative abundances in these relationships.LocationGlobal.Time periodRecent.Major taxa studiedTrees.MethodsWe integrated species abundance records from worldwide forest inventories and associated functional traits (wood density, specific leaf area and seed mass) to obtain a data set of 99,953 to 149,285 plots (depending on the trait) spanning all forested continents. We computed community-weighted and unweighted means of trait values for each plot and related them to three broad environmental gradients and their interactions (energy availability, precipitation and soil properties) at two scales (global and biomes).ResultsOur models explained up to 60% of the variance in trait distribution. At global scale, the energy gradient had the strongest influence on traits. However, within-biome models revealed different relationships among biomes. Notably, the functional composition of tropical forests was more influenced by precipitation and soil properties than energy availability, whereas temperate forests showed the opposite pattern. Depending on the trait studied, response to gradients was more variable and proportionally weaker in boreal forests. Community unweighted means were better predicted than weighted means for almost all models.Main conclusionsWorldwide, trees require a large amount of energy (following latitude) to produce dense wood and seeds, while leaves with large surface to weight ratios are concentrated in temperate forests. However, patterns of functional composition within-biome differ from global patterns due to biome specificities such as the presence of conifers or unique combinations of climatic and soil properties. We recommend assessing the sensitivity of tree functional traits to environmental changes in their geographic context. Furthermore, at a given site, the distribution of tree functional traits appears to be driven more by species presence than species abundance.

AimEcological and anthropogenic factors shift the abundances of dominant and rare tree species within local forest communities, thus affecting species composition and ecosystem functioning. To inform forest and conservation management it is important to understand the drivers of dominance and rarity in local tree communities. We answer the following research questions: (1) What are the patterns of dominance and rarity in tree communities? (2) Which ecological and anthropogenic factors predict these patterns? And (3) what is the extinction risk of locally dominant and rare tree species?LocationGlobal.Time period1990-2017.Major taxa studiedTrees.MethodsWe used 1.2 million forest plots and quantified local tree dominance as the relative plot basal area of the single most dominant species and local rarity as the percentage of species that contribute together to the least 10% of plot basal area. We mapped global community dominance and rarity using machine learning models and evaluated the ecological and anthropogenic predictors with linear models. Extinction risk, for example threatened status, of geographically widespread dominant and rare species was evaluated.ResultsCommunity dominance and rarity show contrasting latitudinal trends, with boreal forests having high levels of dominance and tropical forests having high levels of rarity. Increasing annual precipitation reduces community dominance, probably because precipitation is related to an increase in tree density and richness. Additionally, stand age is positively related to community dominance, due to stem diameter increase of the most dominant species. Surprisingly, we find that locally dominant and rare species, which are geographically widespread in our data, have an equally high rate of elevated extinction due to declining populations through large-scale land degradation.Main conclusionsBy linking patterns and predictors of community dominance and rarity to extinction risk, our results suggest that also widespread species should be considered in large-scale management and conservation practices.

To reduce the potential threat of soil loss due to ephemeral gullies, it is crucial to adopt Best Management Practices (BMPs) that prevent damage to landscapes by reducing sediments load. The current research evaluated the impact of five BMPs, including cover crops, grassed waterways, no-till, conservation tillage, and riparian buffer strips for reduction of sediment load from sheet/rill, and ephemeral gully erosion in an agricultural watershed in Southern Ontario, Canada. The study aimed to automatically calibrate AnnAGNPS using genetic algorithm and the most sensitive parameters of the model identified using a combination of Latin Hypercube Sampling (LHS) and One-At-a-Time (OAT) approach. It also utilized the calibrated model to simulate the effectiveness of BMPs in reducing the average seasonal and annual sediment loads from both sources of erosion (sheet/rill, and ephemeral gully) to determine the most effective practices. Riparian buffer strips were consistently successful in decreasing average seasonal sediment load of sheet/rill erosion, with an average reduction efficiency of 72 % in Spring, 64 % in Summer, 65 % in Fall, and 76 % in Winter. In terms of reducing average seasonal sediment load from ephemeral gully erosion, grassed waterways proved to be the most effective BMPs. They showed efficiency of 90 % in Spring; 83 % in Summer; 79 % in Fall; and 75 % in Winter. Considering the average annual sediment load, riparian buffer strips were consistently successful in decreasing average annual sediment load of sheet/rill erosion, with 69% reduction efficiency. Similarly, grassed waterways were the most effective BMPs for reducing average annual sediment load of ephemeral gully erosion, with an efficiency of 81 %. Additionally, grassed waterways were found to be the most efficient BMPs for reducing average annual total sediment load with reduction efficiency of 71 %. These results demonstrate the importance of implementing effective BMPs to address ephemeral gully erosion in watersheds where ephemeral gullies are the main source of erosion.

The density of wood is a key indicator of the carbon investment strategies of trees, impacting productivity and carbon storage. Despite its importance, the global variation in wood density and its environmental controls remain poorly understood, preventing accurate predictions of global forest carbon stocks. Here we analyse information from 1.1million forest inventory plots alongside wood density data from 10,703 tree species to create a spatially explicit understanding of the global wood density distribution and its drivers. Our findings reveal a pronounced latitudinal gradient, with wood in tropical forests being up to 30% denser than that in boreal forests. In both angiosperms and gymnosperms, hydrothermal conditions represented by annual mean temperature and soil moisture emerged as the primary factors influencing the variation in wood density globally. This indicates similar environmental filters and evolutionary adaptations among distinct plant groups, underscoring the essential role of abiotic factors in determining wood density in forest ecosystems. Additionally, our study highlights the prominent role of disturbance, such as human modification and fire risk, in influencing wood density at more local scales. Factoring in the spatial variation of wood density notably changes the estimates of forest carbon stocks, leading to differences of up to 21% within biomes. Therefore, our research contributes to a deeper understanding of terrestrial biomass distribution and how environmental changes and disturbances impact forest ecosystems.

The density of wood is a key indicator of trees’ carbon investment strategies, impacting productivity and carbon storage. Despite its importance, the global variation in wood density and its environmental controls remain poorly understood, preventing accurate predictions of global forest carbon stocks. Here, we analyze information from 1.1 million forest inventory plots alongside wood density data from 10,703 tree species to create a spatially-explicit understanding of the global wood density distribution and its drivers. Our findings reveal a pronounced latitudinal gradient, with wood in tropical dry forests being up to twice as dense as that in boreal forests. In both angiosperms and gymnosperms, temperature and water availability emerged as the primary factors influencing the variation in wood density globally. This indicates similar environmental filters and evolutionary adaptations among distinct plant groups, underscoring the essential role of abiotic factors in determining wood density in forest ecosystems. Additionally, our study highlights the prominent role of disturbance, such as human modification and fire risk, in influencing wood density at more local scales. Factoring in the spatial variation of wood density notably changes the estimates of forest carbon stocks, leading to differences of up to 21% within biomes. Therefore, our research contributes to a deeper understanding of terrestrial biomass distribution and how environmental changes and disturbances impact forest ecosystems.

In Canada's Arctic tundra region, permafrost is continuous, and the landscape is rich in patterned features. Polygonal terrain, which includes both high- and low-centered features and their wet trenches below, is considered to be high-latitude wetlands in the continuous permafrost region. These prominent hydrological features retain and transport water within widespread ice-wedge networks and govern many ecosystem dynamics. Due to the meter-scale spatial gradients of these processes, mapping of polygonal wetland networks necessitates high-resolution imagery. To date, most studies have used optical imagery for this task; however, these sensors are affected by cloud cover and polar darkness, limiting image availability and repeatability. Thus, our overall objective was to evaluate high-resolution hybrid compact polarimetric (HCP) imagery from the recently launched Radarsat Constellation Mission (RCM), in fusion with ArcticDEM topographic data, for Arctic landscape mapping with a focus on polygonal wetlands. RCM's 5 m Stripmap beam mode, which has yet to be studied for such a task, represents an innovative HCP synthetic aperture radar (SAR) data source since it allows for polarimetric decomposition methods, despite being a dual-pol system. Within this study, we present a seven-input channel Convolutional Neural Network (CNN) model for the classification of ice-wedge dominated landscapes. A range of model hyperparameters as well as the effect of SAR speckle filtering on classification accuracy, have been examined. The optimized CNN achieved a high classification accuracy (0.931 mean Intersection Over Union; mIOU) for three semantic classes representative of the study area, namely polygonal wetlands, open water, and uplands. These results were superior in comparison to a benchmark machine learning (ML) Random Forest (RF) algorithm, thus demonstrating the proposed CNN's potential for regional-scale permafrost feature mapping. Notably, the optimal CNN architecture used unfiltered SAR data as input, underscoring the importance of spatial resolution when classifying polygonal terrain with deep learning (DL). These findings have important implications regarding the design, tuning, and sensitivity of CNN algorithms, and on the efficacy of HCP SAR, for mapping Arctic regions.

Ephemeral gully erosion is one of the main sources of soil loss from agricultural landscapes. Various tools including predictive models with machine learning (ML) algorithms have shown promise to identify susceptible areas. However, ML models have two limitations: (1) a trained model in one area may not be applicable in another area and (2) their application for susceptibility mapping of ephemeral gullies at large-scale areas presents a challenge due to the small size of these features and the need for digitization of all gullies for accurate susceptibility mapping. To overcome these limitations, a novel approach was introduced in the current study for comprehensive validation of ML models and prepare a susceptibility map of ephemeral gullies using an areal transfer of calibration–validation relations. Five ML models were evaluated in Northern Lake Erie Basin as a large-scale region. First, the region was divided into three zones based on the most effective factors of gully formation, and a total of eight watersheds were selected in Zone 1 and Zone 2 (hereafter study area). Zone 3 was not considered, because no gullies were observed in this zone. All the ML models were compared using a new validation approach, including local (trained and validated in the same area) and transferred (trained in one area and tested in other areas). Results showed that random forest (RF) was the most accurate local model in both Zone 1 (accuracy = 0.8833, AUC = 0.8830, sensitivity = 0.9239, and specificity = 0.8537) and Zone 2 (accuracy = 0.8606, AUC = 0.8608, sensitivity = 0.8987, and specificity = 0.8381), while gradient boosting decision tree (GBDT) was the most accurate transferred model (accuracy = 0.7298, AUC = 0.7297, and sensitivity = 0.7826). From the results of the current study, it can be concluded that (1) zonation technique supports the prediction of ephemeral gullies by dividing the study area into the small zones that carry similar topographical and morphological characteristics and (2) the local-transferred validation technique is a helpful method for finding the ML model that can be trained in a small watershed and scale up to the larger area without further calibration.

1. Biodiversity is an important component of natural ecosystems, with higher species richness often correlating with an increase in ecosystem productivity. Yet, this relationship varies substantially across environments, typically becoming less pronounced at high levels of species richness. However, species richness alone cannot reflect all important properties of a community, including community evenness, which may mediate the relationship between biodiversity and productivity. If the evenness of a community correlates negatively with richness across forests globally, then a greater number of species may not always increase overall diversity and productivity of the system. Theoretical work and local empirical studies have shown that the effect of evenness on ecosystem functioning may be especially strong at high richness levels, yet the consistency of this remains untested at a global scale.2. Here, we used a dataset of forests from across the globe, which includes composition, biomass accumulation and net primary productivity, to explore whether productivity correlates with community evenness and richness in a way that evenness appears to buffer the effect of richness. Specifically, we evaluated whether low levels of evenness in speciose communities correlate with the attenuation of the richness-productivity relationship.3. We found that tree species richness and evenness are negatively correlated across forests globally, with highly speciose forests typically comprising a few dominant and many rare species. Furthermore, we found that the correlation between diversity and productivity changes with evenness: at low richness, uneven communities are more productive, while at high richness, even communities are more productive.4. Synthesis. Collectively, these results demonstrate that evenness is an integral component of the relationship between biodiversity and productivity, and that the attenuating effect of richness on forest productivity might be partly explained by low evenness in speciose communities. Productivity generally increases with species richness, until reduced evenness limits the overall increases in community diversity. Our research suggests that evenness is a fundamental component of biodiversity-ecosystem function relationships, and is of critical importance for guiding conservation and sustainable ecosystem management decisions.

Understanding what controls global leaf type variation in trees is crucial for comprehending their role in terrestrial ecosystems, including carbon, water and nutrient dynamics. Yet our understanding of the factors influencing forest leaf types remains incomplete, leaving us uncertain about the global proportions of needle-leaved, broadleaved, evergreen and deciduous trees. To address these gaps, we conducted a global, ground-sourced assessment of forest leaf-type variation by integrating forest inventory data with comprehensive leaf form (broadleaf vs needle-leaf) and habit (evergreen vs deciduous) records. We found that global variation in leaf habit is primarily driven by isothermality and soil characteristics, while leaf form is predominantly driven by temperature. Given these relationships, we estimate that 38% of global tree individuals are needle-leaved evergreen, 29% are broadleaved evergreen, 27% are broadleaved deciduous and 5% are needle-leaved deciduous. The aboveground biomass distribution among these tree types is approximately 21% (126.4 Gt), 54% (335.7 Gt), 22% (136.2 Gt) and 3% (18.7 Gt), respectively. We further project that, depending on future emissions pathways, 17–34% of forested areas will experience climate conditions by the end of the century that currently support a different forest type, highlighting the intensification of climatic stress on existing forests. By quantifying the distribution of tree leaf types and their corresponding biomass, and identifying regions where climate change will exert greatest pressure on current leaf types, our results can help improve predictions of future terrestrial ecosystem functioning and carbon cycling.