Explore the vast range of titles available.

AIM: Functional traits of organisms are key to understanding and predicting biodiversity and ecological change, which motivates continuous collection of traits and their integration into global databases. Such trait matrices are inherently sparse, severely limiting their usefulness for further analyses. On the other hand, traits are characterized by the phylogenetic trait signal, trait–trait correlations and environmental constraints, all of which provide information that could be used to statistically fill gaps. We propose the application of probabilistic models which, for the first time, utilize all three characteristics to fill gaps in trait databases and predict trait values at larger spatial scales. INNOVATION: For this purpose we introduce BHPMF, a hierarchical Bayesian extension of probabilistic matrix factorization (PMF). PMF is a machine learning technique which exploits the correlation structure of sparse matrices to impute missing entries. BHPMF additionally utilizes the taxonomic hierarchy for trait prediction and provides uncertainty estimates for each imputation. In combination with multiple regression against environmental information, BHPMF allows for extrapolation from point measurements to larger spatial scales. We demonstrate the applicability of BHPMF in ecological contexts, using different plant functional trait datasets, also comparing results to taking the species mean and PMF. MAIN CONCLUSIONS: Sensitivity analyses validate the robustness and accuracy of BHPMF: our method captures the correlation structure of the trait matrix as well as the phylogenetic trait signal – also for extremely sparse trait matrices – and provides a robust measure of confidence in prediction accuracy for each missing entry. The combination of BHPMF with environmental constraints provides a promising concept to extrapolate traits beyond sampled regions, accounting for intraspecific trait variability. We conclude that BHPMF and its derivatives have a high potential to support future trait‐based research in macroecology and functional biogeography.

Soil salinity is a major ecological challenge that affects agricultural productivity, posed significant challenges on the ecological system, especially in the deltaic region vulnerable to human alterations and sea level rise. Assessing agricultural areas impacted by soil salinity change is very important to support decision-makers or planners in sustainable land use planning. To overcome limitations in current spatial extrapolation methods for a reliable prediction of salinity trends across extensive river deltas, an advanced synthesis approach was developed with the use of machine learning (ML) particularly appropriate to account for a multitude of factors representing land cover conditions, processes, and interactions. This study aims to: (i) address the extrapolation challenge in ML-based soil salinity mapping, and (ii) predict land cover changes due to soil salinity. The Mekong River Delta (MRD) and Red River Delta (RRD) were selected as case studies. A hybrid ML approach and land change modeler were used to analyze 39 contributing factors. To resolve the spatial extrapolation issue in soil salinity monitoring, we used 109 salinity-affected locations in the MRD and 72 in the RRD. Land cover data from 2000 and 2023, along with salinity maps, were used to project the 2050 land cover. Multiple ML models were used to cross-verify and obtain robust results. All models achieved R2 scores above 0.85, with the best model exceeding 0.93, demonstrating high predictive performance. Among the models, XGR-particle swarm optimization achieved the highest accuracy (R2 = 0.939), followed closely by XGR-fennec fox optimization, XGR-coati optimization algorithm (R2 = 0.932), and XGR-osprey optimization algorithm (R2 = 0.921), respectively, highlighting the robustness of optimization-enhanced XGBoost models. Future projections show that cropland will decline from 67% of the area (in 2000) to 64% (2023) and about 60% (2050) under the influence of salinity, with approximately 41 km2 of cropland converted to aquaculture by 2050, mostly in high-salinity coastal zones. This study develops a powerful synthesis framework to address the problem of spatial extrapolation challenges related to natural hazard mapping in general and soil salinity mapping in particular, based on ML and on accurate prediction of land cover/land use change under effects of soil salinity in the context of climate change. Results from the synthesis approach help accurately identify areas affected by salinity intrusion, useful for the development of effective solutions in space and time towards the goal of sustainable development.

There is an increasing need for ecosystem-level distribution models (EDMs) and a better understanding of which factors affect their quality. We investigated how the performance and transferability of EDMs are influenced by 1) the choice of predictors and 2) model complexity. We modelled the distribution of 15 pre-classified ecosystem types in Norway using 252 predictors gridded to 100 × 100 m resolution. The ecosystem types are major types in the ‘Nature in Norway' system mainly defined by rule-based criteria such as whether soil or specific functional groups (e.g. trees) are present. The predictors were categorised into four groups, of which three represented proxies for natural, anthropogenic, or terrain processes (‘ecological predictors') and one represented spectral and structural characteristics of the surface observable from above (‘surface predictors'). Models were generated for five levels of model complexity. Model performance and transferability were evaluated with data collected independently of the training data. We found that 1) models trained with surface predictors only performed considerably better and were more transferable than models trained with ecological predictors, and 2) model performance increased with model complexity, levelling off from approximately 10 parameters and reaching a peak at approximately 20 parameters, while model transferability decreased with model complexity. Our findings suggest that surface predictors enhance EDM performance and transferability, most likely because they represent discernible surface characteristics of the ecosystem types. A poor match between the rule-based criteria that define the ecosystem types and the ecological predictors, which represent ecological processes, is a plausible explanation for why surface predictors better predict the distribution of ecosystem types. Our results indicate that, in most cases, the same models are not well suited for contrasting purposes, such as predicting where ecosystems are and explaining why they are there.

Meteorological station measurements are an important source of information for understanding the weather and its association with risk, and are vital in quantifying climate change. However, such data tend to lack spatial coverage and are often plagued with flaws such as erroneous outliers and missing values. Alternative meteorological data exist in the form of climate model output that have better spatial coverage, at the expense of bias. We propose a probabilistic framework to integrate temperature measurements with climate model (reanalysis) data, in a way that allows for biases and erroneous outliers, while enabling prediction at any spatial resolution. The approach is Bayesian which facilitates uncertainty quantification and simulation based inference, as illustrated by application to two countries from the Middle East and North Africa region, an important climate change hotspot. We demonstrate the use of the model in: identifying outliers, imputing missing values, non-linear bias correction, downscaling and aggregation to any given spatial configuration.

Accurate estimation of forest aboveground biomass (AGB) and understanding its ecological drivers are vital for carbon monitoring and sustainable forest management. However, AGB estimation using remote sensing is hindered by signal saturation in high-biomass areas and insufficient attention to ecological structural factors. Focusing on Guangdong Province, this study proposes a novel approach that spatially extrapolates airborne LiDAR-derived Forest structural parameters and integrates them with Sentinel-1/2 data to construct an AGB prediction model. Results show that incorporating structural parameters significantly reduces saturation effects, improving prediction accuracy and AGB maximum range in high-AGB regions (R2 from 0.724 to 0.811; RMSE = 10.64 Mg/ha; max AGB > 180 Mg/ha). Using multi-scale geographically weighted regression (MGWR), we further examined the spatial influence of forest type, age structure, and species mixture. Forest age showed a strong positive correlation with AGB in over 95% of the area, particularly in mountainous and hilly regions (coefficients up to 1.23). Species mixture had positive effects in 87.7% of the region, especially in the north and parts of the south. Natural forests consistently exhibited higher AGB than plantations, with differences amplifying at later successional stages. Highly mixed natural forests showed faster biomass accumulation and higher steady-state AGB, highlighting the regulatory role of structural complexity and successional maturity. This study not only mitigates remote sensing saturation issues but also deepens understanding of spatial and ecological drivers of AGB, offering theoretical and technical support for targeted carbon stock assessment and forest management strategies.

High-resolution wind datasets are crucial for ultra-short-term wind forecasting. Penetration of WT installations near urban areas that are constantly changing will motivate researchers to understand how to adapt their models to terrain changes to reduce forecasting errors. Although CFD modelling is not widely used for ultra-short-term forecasting purposes, it can overcome such difficulties. In this research, we will spatially extrapolate vertical profile LIDAR wind measurements into a 3D wind velocity field over a large and relatively complex terrain with the use of stationary CFD simulations. The extrapolated field is validated with measurements at a hub height of three WTs located in the area. The accuracy of the model increases with height because of the terrain anomalies and turbulence effects. The maximum MAE of wind velocity at WT hub height is 0.81 m/s, and MAPE is 7.98%. Our model remains accurate even with great simplifications and scarce measurements for the complex terrain conditions of our case study. The models’ performance under such circumstances establishes it as a promising tool for the evolution of ultra-short-term forecasting as well as for the evaluation of new WT installations by providing valuable data for all models.

The French Moss Survey employs forest mosses as indicators to monitor the deposition of atmospheric substances, notably focusing on cadmium (Cd), a known carcinogenic and contributor to respiratory illnesses. This comprehensive study encompasses 55 variables to understand Cd accumulation in terrestrial mosses in France. These variables include moss species, tree cover, biogeographical markers, land use area, proximity to road and rail networks, soil concentration of Cd and atmospheric concentration and deposition of Cd using a physical model. The response variable undergoes a complementary log–log transformation to constrain prediction values within the maximum Cd content in mosses. We have built a regression model to improve predictions, considering the impacts of covariates in France. This model retains biogeographical effects, leading to data segmentation into four distinct biogeographical zones: Atlantic, Continental, Mediterranean and Alpine. Subsequently, zone-specific regression models are explored to refine predictions and consider the impacts of covariates specific to each region, such as those related to railways and roads of the Mediterranean zone. Our biogeographical models effectively mitigate spatial correlation issues and yield accurate predictions, as evidenced by the leave-one-out cross-validation assessment. Compared to ordinary kriging map, the regression prediction maps highlight the contributions of certain covariates, such as the EMEP atmospheric transport model, to areas with high Cd concentrations. Furthermore, these maps exhibit new areas with high (resp. low) Cd concentrations due to high (resp. low) values of the covariates.

Shallow-water depth estimates from airborne lidar data might be improved by using sounding attribute data (SAD) and ocean geomorphometry derived from lidar soundings. Moreover, an accurate derivation of geomorphometry would be beneficial to other applications. The SAD examined here included routinely collected variables such as sounding intensity and fore/aft scan direction. Ocean-floor geomorphometry was described by slope, orientation, and pulse orthogonality that were derived from the depth estimates of bathymetry soundings using spatial extrapolation and interpolation. Four data case studies (CSs) located near Key West, Florida (United States) were the testbed for this study. To identify bathymetry soundings in lidar point clouds, extreme gradient boosting (XGB) models were fitted for all seven possible combinations of three variable suites—SAD, derived geomorphometry, and sounding depth. R2 values for the best models were between 0.6 and 0.99, and global accuracy values were between 85% and 95%. Lidar depth alone had the strongest relationship to bathymetry for all but the shallowest CS, but the SAD provided demonstrable model improvements for all CSs. The derived geomorphometry variables contained little bathymetric information. Whereas the SAD showed promise for improving the extraction of bathymetry from lidar point clouds, the derived geomorphometry variables do not appear to describe geomorphometry well.

Watersheds are often degraded by human activities, reducing their ability to provide ecosystem functions and services. While governmental agencies have put forward plans for improving watershed health, resources are limited, and choices must be made as to which watersheds to prioritize and what actions to take. Prioritization tools with sufficient specificity, resolution, and automation are needed to guide decisions on restoration and management actions across large scales. To address this need, we developed a set of tools to support the protection of streams and associated riparian habitats across the state of California. We developed and tested watershed condition estimation models based on bioassessment data, used the EPA’s StreamCat dataset to identify stressors, incorporated environmental justice factors and developed reach-specific models to prioritize actions. We applied the prioritization tools statewide and were able to identify 18% of stream reaches that are in good condition but that are most vulnerable to existing stressors and an additional 19% of stream reaches that are degraded and are highest priority for restoration and management. The remaining 63% of stream reaches were prioritized for protection and periodic monitoring or minor remedial actions. The results of this project can help regional stakeholders and agencies prioritize hundreds of millions of dollars being spent to protect, acquire, and restore stream and riparian habitats. The methods are directly transferable by using any regional condition and stress data that can be readily obtained.

Bubble transport of methane from shallow seep sites in the Black Sea west of the Crimea Peninsula between 70 and 112 m water depth has been studied by extrapolation of results gained through different hydroacoustic methods and direct sampling. Ship‐based hydroacoustic echo sounders can locate bubble releasing seep sites very precisely and facilitate their correlation with geological or other features at the seafloor. Here, the backscatter strength of a multibeam system was integrated with single‐beam data to estimate the amount of seeps/m2 for different backscatter intensities, resulting in 2709 vents in total. Direct flux measurements by submersible revealed methane fluxes from individual vents of 0.32–0.85 l/min or 14.5–37.8 mmol/min at ambient pressure and temperature conditions. A conservative estimate of 30 mmol/min per site was used to estimate the flux into the water to be 1219–1355 mmol/s. The flux to the atmosphere was calculated by applying a bubble dissolution model taking release depth, temperature, gas composition, and bubble size spectra into account. The flux into the atmosphere (3930–4533 mol/d) or into the mixed layer (6186–6899 mol/d) from the 21.8 km2 large study area is three times higher than independently measured fluxes of dissolved methane for the same area using geochemical methods (1030–2495 mol/d). The amount of methane dissolving in the mixed layer is 2256–2366 mol/d. This close match shows that the hydroacoustic approach for extrapolating the number of seeps/m2 and the applied bubble dissolution model are suitable to extrapolate methane fluxes over larger areas.