Search

• The selection of satellite sensor and radar system parameters greatly affected the model output. • The model was improved when more polarizations, orbital directions and frequencies were involved. • Models built using multiband radar datasets achieved a comparable accuracy to models based on optical data. • The model built by fusing SAR-optical data achieved better results than the model without SAR data. • The development of this continental-scale DSM work largely benefits from GEE. The advent of cloud computing platforms (e.g., Google Earth Engine (GEE)) and the massive amounts of optical and radar Earth Observation (EO) data hosted by these platforms present new opportunities for mapping soil pH at large scales. However, existing studies generally lack consensus on the effects of satellite and radar sensor parameters on GEE-based soil pH prediction models. In this study, we assessed the suitability of long-term radar (C-band Sentinel-1 and L-band PALSAR-1/2) and optical (Sentinel-2) EO data on GEE for the digital mapping of soil pH on a continental (Europe) scale and determined the most appropriate radar sensor parameters. Thirteen scenarios with different data configurations were simulated and combined with the 2018 LUCAS soil database and two machine learners (boosted regression trees and extreme gradient boosting) to develop soil prediction models. Results showed that the selection of modeling techniques, satellite sensors and radar system parameters largely affected the model output. Models involving a single polarization mode of PALSAR-1/2 data performed the worst (RPD = 1.24). Models based on Sentinel-1 data performed better than those built using PALSAR-1/2 data. The model performance was improved when a model involved more polarization bands, orbital directions, and band frequencies. The multiband model built using the two radar datasets achieved a comparable accuracy to the model based on optical data. Moreover, the model that fused radar-optical data achieved better results, with RPD values of 1.56 and 1.46 for the models with and without radar data, respectively; its performance was comparable to that of models built with commonly used variables (topography and climate). The analysis of importance indicated that long-term optical and radar EO data on GEE were important in our model. The modelling of soil pH at the continental scale largely benefits from GEE. The predicted maps exhibited strong spatial heterogeneity among different biogeographic regions, with similar spatial patterns under different modelling scenarios. [ABSTRACT FROM AUTHOR]

[Display omitted] • Establishment of an object-oriented CART model for high-resolution identification of riverine wetlands. • Localization of a Pressure-State-Response model for riverine wetland ecosystem health assessment. • Urbanization tends to more fragmentation and less connectivity of riverine wetlands landscape. • Safeguarding existing wetlands should be paid more attention in riverine urban agglomerations. Riverine wetland is one of the important cityscapes along rivers, featuring powerful eco-hydrological regulations in safeguarding urban security and serving its quality. Over the past several decades, intensified climate change, together with the upgraded human activities, have deeply disturbed riverine wetland worldwide and caused variations in the amount and pattern, which might lead to negative effects on wetland ecosystem health (WEH), thus threaten the sustainability of the riverine urban agglomerations. To better understand the mechanism of the above effectiveness, the Riverine Urban Agglomerations along the Yellow River in China's Ningxia Hui Autonomous Region (RUAN) was taken as an example in the present study. First, by establishing an object-oriented remote sensing image classification system based on Classification and Regression Tree (CART), riverine wetland distribution in years of 2000, 2009 and 2018 were determined for a two-Stage (ST-I: 2000–2009, ST-II: 2009–2018) comparative research. Second, the transition matrix and landscape index were used to measure the spatiotemporal dynamics of the riverine wetland in the two stages. Third, a Pressure-State-Response (P-S-R) model, together with its index system, was constructed to comprehensively assess the WEH in there. Results revealed that: (1) Wetlands in RUAN are dominated by the artificial ones, presenting an overall increase in area during the statistical period, varying with different trends in the two stages. Conversion between the wetlands and the non-wetlands were found frequent during the urbanization, leading to remarkable changes in wetland patterns in space. (2) Rivers are important and basic in forming riverine landscape in RUAN, the natural wetlands increased during the statistical period. In general, wetland patches and diversity increased, shapes of that became homogenized. Accordingly, the aggregation of wetlands decreased and the fragmentation of that worsened. (3) Riverine wetlands in most of RUAN experienced an increase of external pressure and a deterioration of state, showing an overall degradation. Meanwhile, the WEH was determined more by the fundamental state of itself, the external disturbance seemed function less. Above findings confirmed the vulnerability of riverine wetlands during urbanization in an arid circumstance. It is worthy of strengthening the protection of existing wetlands and minimizing or eliminating the conversion, to ensure the WEH and serve the sustainability of riverine urban agglomerations. [ABSTRACT FROM AUTHOR]

[Display omitted] • From 2000 to 2020, green space increased more than did blue space in central Nanjing. • Fringe-type expansion of urban green–blue space (UGBS) was most common in central Nanjing. • Average annual solar radiation, DEM, slope, and aspect are natural factors influencing the evolution of UGBSs. • Population density is a socio-economic factor affecting the evolution of UGBSs. Urban green–blue space (UGBS) plays a crucial role in Chinese cities, occupying more than half of the built-up areas. Understanding the spatiotemporal dynamics and driving factors of UGBS in central Nanjing during rapid urbanization is important for studying UGBS evolution in high-density urban-area environments, providing a fundamental basis for related research and policy formulation. This study utilized remote sensing images to estimate the extent of UGBS from 2000 to 2020 and analyzed the evolution using the intensity of area change, a land transfer matrix, the multi-order adjacency index, and a boosted regression tree model. This study revealed the spatiotemporal dynamics and driving factors of UGBSs in central Nanjing from 2000 to 2020: (1) between 2000 and 2020, there was a 4.0% net increase in the total area of UGBSs in central Nanjing. (2) UGBS primarily showed marginal-type expansion, accompanied by proximity-type expansion. Spatially, new green–blue patches tended to cluster around large pre-existing natural green–blue patches. Furthermore, the expansion of UGBS shifted from the urban periphery to the urban interior. (3) Natural factors exerted a greater influence (88.3%) than socioeconomic factors (11.7%) on the expansion of UGBS. Average annual radiation (43.4%), elevation (16.5%), slope (15%), aspect (13.4%), and population density (2.8%) were the primary driving factors. Our results contribute to understanding changes in UGBS and the driving factors behind these changes, thereby providing support for the future planning, protection, and sustainable development of UGBS. [ABSTRACT FROM AUTHOR]

• A database on the distribution of Iberian peatlands has been created. • Land use transitions between 1990 and 2018 were assessed. • The expansion of agricultural and artificial land uses was observed. • Lowland peatlands showed higher susceptibility to land use change. • Elevation and agricultural density are related to variation in changed area. Iberian peatlands have been severely affected by land use and land cover (LULC) changes. Despite these pressures, some peatlands persist in the region, although their susceptibility to LULC change remains poorly understood. This study presents the most detailed and extensive distribution data for Iberian peatlands to date and analyzes the dynamics and drivers of LULC in Iberian peatlands and their surrounding areas. We compiled peatland records from various sources and used Corine Land Cover Change layers to determine LULC shifts in their surrounding areas for 1990, 2000, 2006, 2012, and 2018. Environmental and socioeconomic variables were used to create Boosted Regression Tree models explaining spatial variations in the mean percentage of changed area. Analysis of 270 peatland locations in the Iberian Peninsula revealed that forests and seminatural areas constituted over 80% of the peatland's surroundings. Agricultural areas expanded the most, except between 2006 and 2012 when the artificial areas showed more gains. While most surrounding areas experienced an average change of 0%-9.51% of the total area, between 1990 and 2018, lowland peatlands (littoral and sublittoral) suffered more intense changes (9.51% to 38.43%). Our models showed that only elevation and agricultural area density were relevant predictors of the spatial distribution of changes. Upland Iberian peatlands showed lower susceptibility to LULC change, while lowland peatlands underwent remarkable transformations. This study substantially expands previous knowledge about the distribution and conservation needs of these ecosystems in the Iberian Peninsula, especially those in littoral and sublittoral lowlands. [ABSTRACT FROM AUTHOR]

• Daily reference evapotranspiration forecasts via using climate-derived inputs. • Design novel multivariate variational mode decomposition (MVMD) hybridized with boost regression tree (BRT). • Comparative assessment between MVMD and multivariate empirical mode decomposition (MEMD) for the first time. • The hybrid MVMD based models better in accuracy than MEMD based models. • The novel hybrid model provides significant implications for reference evapotranspiration. Reference evapotranspiration (ET o) is a vital climate parameter affecting plants' water use. ET o can generate large deficits in soil moisture and runoff in different regions and seasons, leading to uncertainties in drought warning systems. A novel multivariate variational mode decomposition integrated with a boosted regression tree (i.e., MVMD-BRT) is constructed to forecast daily ET o. Firstly, the correlation matrix based on cross-correlation was computed to investigate the significant input predictor lags of daily ET o. Secondly, the MVMD technique decomposes the significant input lags into signals called intrinsic mode functions (IMFs). Thirdly, the IMFs were then employed in the BRT to build the MVMD-BRT model for daily ETo forecasting. A comparative assessment of MVMD against multivariate empirical mode decomposition (MEMD) was also performed on the same lines to develop the MEMD-BRT model. The MVMD-BRT model is compared against the random forest (RF) and hybrid MVMD-RF, MEMD-RF, extreme learning machine (ELM), and hybrid MVMD-ELM, MEMD-ELM, and cascaded feedforward neural network (CFNN) along with its hybrid MVMD-CFNN models for two stations in Queensland, Australia using a set of goodness-of-fit metrics. The results prove that the MVMD-BRT provide accurate daily ET o forecasting against the benchmark models. The MVMD-BRT model yielded the highest accuracy in terms of (WI E = 0.9070 , NS E = 0.8421, LM E = 0.6529, KGE = 0.8792) and (WI E = 0.8966 , NS E = 0.8396, LM E = 0.6521, KGE = 0.8803) for Brisbane and Gympie stations against the comparing models. [ABSTRACT FROM AUTHOR]

• The environmental niche model was used to explain and predict the potential distribution of fishing effort. • Environmental preferences of squid-jigging fleet could be characterized by several environmental variables. • Potential fishing areas of squid-jigging fleet could be identified in other sea areas. • The selected variable and its relative contribution vary in each month. • The BRT shows specific nonlinear response thresholds for different driving factors. The vast number of fishing fleets and limited regulatory capacity pose challenges to the effective monitoring and management of marine fishery resources. To understand the fishing characteristics of spatial distribution for squid-jigging fleets in the North Pacific Ocean (NPO), satellite-based automatic identification system (AIS) data during 2016 ∼ 2020 and marine environmental data were applied to explore the drivers of fishing activities based on the boosted regression tree (BRT) and generalized additive model (GAM). The results showed that the BRT model has better performance than the GAM. The spatial distribution of squid-jigging fleets has a significant seasonality. Specifically, the spatial distribution of Fishing effort (FE) gradually shifted to the northwest from May to September, while gradually shifted to the southwest from October to November. The favourable area for focusing FE was the 150 ° E ∼ 165 ° E, 40 ° N ∼ 44 ° N. The latitudinal distribution of FE was not wide and the longitudinal distribution was long. Variations in the marine environment constantly influenced the FE of the squid-jigging fleets. Overall, the influence of sea surface temperature ( T 0 ) on FE was most significant. The favourable range of T 0 for FE was 12 ∼ 20 ° C. Water temperature at the 300 m depth ( T 300 ) was also an important factor through the fishing season. The favourable range of T 300 for FE was 3 ∼ 6 ° C. There were obvious differences in the influence of other environmental factors on FE in different months. Different environmental variables drive the spatial distribution of fishing effort in different months. This study may help to scientifically and effectively guide fishery management and sustainable development by evaluating the spatial variations in fishing activity. [ABSTRACT FROM AUTHOR]

[Display omitted] • Buildings are crucial landscape elements that impact spontaneous plant richness. • BCR, MAH, SCD, FAR, ED_G, and GCR are the key influencing metrics. • These metrics have nonlinear correlations with spontaneous species richness. • Native/non-native species richness respond differently to building/green space patterns. The impact of landscape patterns on urban plant diversity has received significant attention; however, previous studies have primarily focused on two-dimensional (2D) patterns. Limited investigations have been conducted to explore the effects of three-dimensional (3D) landscape patterns on urban plant diversity. Spontaneous plants are ideal objects for studying the response of plant diversity to urban environments due to their independence from intentional human inputs. Buildings, as a crucial symbol of urbanization, exhibit strong 3D characteristics in urban landscapes. Green spaces, the most important source of plant propagules in cities, are closely related to spontaneous plant diversity in urban areas. Therefore, we examined how 2D/3D building and green space patterns influence urban spontaneous plant species richness and compared the responses between different categories (native and non-native). We surveyed spontaneous plants in the built-up areas of Shenzhen, a highly urbanized region situated along China's southern coast. Our survey recorded 278 species belonging to 208 genera and 77 families. Results from boosted regression tree models (BRT) indicated significant correlations between spontaneous plant species richness and various metrics, including building coverage ratio (BCR), mean building height (MAH), green space coverage ratio (GCR), edge density of green spaces (ED_G), spatial congestion degree (SCD), and floor area ratio (FAR). These metrics exhibited complex nonlinear correlations with the species richness of spontaneous plants. Furthermore, native and non-native species richness responded differently to building and green space patterns; while native species richness was equally affected by green space and building patterns, non-native species richness was more strongly impacted by building patterns than green space ones. Our study quantified the nonlinear relationships between urban spontaneous plant species richness and 2D/3D green space and building patterns, demonstrating that optimizing green space and building patterns can be a practical approach for managing and promoting spontaneous plant diversity in highly urbanized areas. [ABSTRACT FROM AUTHOR]

• Ratio estimators are simple to apply compared to non-linear regression. • Ratio estimators promote the natural additivity of biomass parts. • Stratification corrected the heteroscedasticity problem. • Stratification promotes accuracy gains. Researchers on forest biomass seek to develop or find methods of simple application, whose estimators reveals the best performance and provide the additivity of the biomass parts. Tree biomass normally is obtained by scaling tree parts (stem, branches, and leaves) separately. When the biomass is modeled by parts of the tree, their sum does not match the total, which means lack of additivity in the tree biomass estimation. We aimed to test and propose a modeling technique for estimating total aboveground biomass, so that consistency between tree parts will be achieved. The techniques 'ratio estimators' and 'weighted apparently unrelated nonlinear regressions (WNSUR)' were tested using a biomass dataset with 387 trees. Data were collected in eight sites located in Paraná and Rio Grande do Sul, Brazil, where information was collected on diameter at 1.30 m aboveground (dbh), heights (total, stem and crown), total biomass aboveground and its parts: stem, branches and leaves. All trees were identified at species level. To adjust the ratio equations, the cylinder volume (v cl), based on the mean square diameter (d qi 2), was used as the independent variable and the total aboveground biomass and the respective parts: stem, branches, and leaves, as dependent variables. The occurrence of heteroscedasticity in the ratio estimators indicated that the data should be stratified into two stages, the first by dbh and the second by the slope of the ratio between the total aboveground biomass and the v cl of each tree. To group the trees in the second stage, a posteriori, a discriminant analysis was applied. The standard errors of the estimate resulted in less than 4% for the ratio estimators, while for the WNSUR they reached values larger than 40% for the total biomass aboveground and its parts. The statistics: bias , MAE , MSE and RMSE showed better performance by ratio estimators when compared to the WNSUR , for the total biomass and its parts except for the bias towards the stem. Ratio estimators naturally provide additivity to biomass parts making it possible to improve the precisions of the estimates. In future work with forest biomass and its parts, more canopy variables should be collected to improve the stratification of trees within the diameter stratum. The total volume of the trees should also be estimated, if possible, via xylometry, to approximate the ratio coefficients of the total aboveground biomass, as much as possible, to the actual values of the specific gravity of the wood. [ABSTRACT FROM AUTHOR]

• Regression trees and classification techniques identified environmental indicators of stream flow at ungagged locations. • Nine out of 77 variables found with higher predictive potential of stream flow. • The predictors were accurate in estimating general annual trends of stream flow. • The nine indicators were less accurate when estimating stream flow by month. The aim of this study is to test a spatially explicit statistical model to identify indicators of natural stream flow using readily available stream, climate and landscape data. Understanding flow behavior of unmonitored streams at different temporal scales using environmental indicators is of great interest considering the logistic constraints of providing comprehensive flow instrumentation for all stream reaches in a region. Our results have applications to assess human impact in watersheds, to study environmental changes in fresh water resources, and in the management of local ecosystem. This study uses classification and regression tree analysis to identify significant explanatory variables for a predictive model of stream flow in semi-arid watersheds of southern California, USA. The study collected 77 variables with 30 years record, for a set of 48 sites, interpolated to create raster files at 30 m spatial resolution. After applying Pearson correlation analyses to eliminate redundant variables, nine variables were found to have strong positive predictive value for estimating stream flow at ungauged sites. Nine prediction rasters portraying spatial variation of stream flow for the study region at three key index months during wet, dry, and average rainfall conditions. The predictive power of the variables was tested and cross validated over a subset of data not included when building the model. Model validation by site at monthly temporal resolution showed mixed results. While some sites where accurately predicted others did not. The comparison of observed vs predicted values by month suggest that this statistically based approach is able to predict the general patterns of stream flow at the regional scale, however it may be inaccurate in estimating actual flow values by month since the models tends to under-predict monthly discharge. [ABSTRACT FROM AUTHOR]

Highlights • Forest recovery indicators may show uneven spatial patterns at landscape scales. • Overall recovery completeness in the Mediterranean forest under natural regeneration averages 72.7% • Recovery completeness was higher for diversity than for structural forest attributes. • Low spatial congruence among recovery indicators may hinder restoration monitoring. Abstract Natural regeneration has been proposed as a cost-effective forest restoration approach for both small and large-scale initiatives. However, attributes for assessing the success of forest restoration through natural regeneration may vary among them in spatial patterns depending on the scale of analysis and on environmental gradients. Here we analysed the spatial patterns of recovery completeness (i.e. how similar attributes in restored forests are to the same attributes in reference forests) in response to environmental factors in a Mediterranean forest landscape of Central Chile. We evaluated (1) forest recovery completeness using basal area (BA), quadratic mean diameter (QMD), adult species density (ASD), adult species richness (ASR), and seedling species richness (SSR); (2) the spatial congruence of recovery completeness estimated by each of these indicators; and (3) the environmental factors potentially shaping these spatial patterns. We used field measurements and geospatial information sources to quantify and predict indicator responses by fitting boosted regression tree models. To assess the spatial congruence of predictions we overlaid high-level recovery completeness values for all indicators. Overall recovery completeness in the study area was 72.7%, suggesting positive prospects for attaining fully restored forests. Recovery completeness had a resulted higher for diversity (92.3%−99.6%) than structural forest attributes (33.5%−76.9%); however, spatial congruence among recovery indicators was low due to the uneven spatial responses of each indicator. The maximum potential spatial congruence was <10%, and was predicted only by two environmental variables (soil bulk density and slope). Our results suggest that low spatial congruence among forest recovery indicators may hinder the monitoring of restoration at large scales. The implications of such divergence in defining restoration success can be enormous given the current global challenge of forest restoration. Although our research was tested in a threatened region of global importance, our results may have wider significance for restoration planning providing cautionary notes and recommendations for the appropriate use of forest recovery indicators when monitoring large-scale restoration projects. [ABSTRACT FROM AUTHOR]

Highlights • Topography is an important driver of diversity and ecosystem functioning. • Fine-scale topography drive woody species richness and hyperdominance. • High topographic heterogeneity induces high richness and hyperdominance. • Stem and biomass hyperdominant species increases along with richness. Abstract Topography is an important driver that determines diversity patterns and ecosystem functioning in tropical forests. However, there are few studies analyzing contrasting topographical conditions on the relative importance of species to ecosystem functioning, mainly on those who have a greater contribution (i.e., hyperdominant species). We aimed to evaluate whether contrasting topographical conditions determine changes in tree species richness, community composition, and the number of stem and biomass hyperdominant species in a Brazilian Atlantic remnant forest fragment. We selected two areas on distinct hillsides with contrasting topographic conditions, at the biological reserve of the Federal University of Viçosa, Minas Gerais state, southeastern Brazil. Each area (100 × 100 m) was sub-divided into 100 plots of 10 × 10 m. From each plot, all trees having diameter at breast height (DBH) ≥ 10 cm were identified to the species level and tagged for measurement. We measured three topographic variables (slope, elevation, and convexity) in each plot, based on the assumption that these variables may affect tree species diversity, species composition, and ecosystem function (aboveground biomass). The AGB of individual stems was calculated in all plots. We performed a multivariate regression tree for estimating topographical heterogeneity in each area. We found that species richness differed significantly between areas. Species richness in the Northeast area (the more topographically heterogeneous one) was 48% higher than that in the Southeast area, which is less topographically heterogeneous. The tree species composition varied considerably between areas, with similar AGB patterns being registered among plots. The number of stem hyperdominants varied significantly between areas. In the Southeast area, only two species out of the 85 recorded (2.38%) accounted for 50% of the number of stems hyperdominants, while in the Northeast area 10 species (7.94%) accounted for 50% of stems hyperdominants. Our results showed that high topographic heterogeneity induces high species richness and that the number of stem and biomass hyperdominant species increase along with richness on a local scale. Based on our results, we presume that biomass hyperdominance can also strongly influence forest ecosystem functioning on a local scale. [ABSTRACT FROM AUTHOR]

Highlights • We partition Italian municipalities into homogeneous groups with similar features. • To form the groups, we use the drivers of separate waste collection. • The quality of local institutions emerges a key variable in this procedure. • Morphology, cultural consumption and the income also play a major role. • The municipalities of Veneto, Campania and Sardinia achieve remarkable results. Abstract The European Commission identifies separate waste collection as a primary factor in its sustainable development strategy. The EU Member States are expected to achieve a separate waste collection rate of 50% by 2020. Although Italy is not very far from such target, it features wide regional variation in the extent of separate collection. Municipal administrations are the key managerial units for the implementation of the environmental policy. By means of the Classification and Regression Tree Analysis, this work assembles Italian municipalities (NUTS-4) into homogenous groups, that share common characteristics with respect to geographic, economic and socio-cultural covariates. Three benchmark regions are identified, i.e. Veneto, Sardinia and much of Campania. The results achieved in these areas depend on the joint action of citizens and institutions and may inspire local-level waste management policies in other areas of the country. [ABSTRACT FROM AUTHOR]

Highlights • We present a Landsat-based method for mapping abandoned land in heterogeneous and fragmented urban landscapes. • The method makes use of NDVI values and decision trees. • Vegetation phenology is accounted for indetermining which period of the growing season fits best for mapping. • Scenes encompassing the beginning of the vegetation growing season enhance classification accuracy. • Monitoring land abandonment and urban sprawl can support evaluation of land-use regulations and planning strategies. Abstract Studies have shown that spatial information on abandoned land could play an important role in urban land management, as land abandonment was proven capable of revealing future trajectories of change. However, mapping land abandonment with traditional methods (e.g., field work, digitization of aerial images) can be time consuming, expensive, and require considerable man-power. In this context, Landsat imagery proves to be a reliable source of data. To our knowledge, the potential of Landsat imagery for mapping abandoned land has not been tested in heterogeneous and highly fragmented urban settings. The aim of our paper is to propose a resource-efficient (i.e., in terms of time and manpower) method for the assessment of land abandonment in areas affected by urban sprawl, by using seasonal time series of Landsat data. Bucharest, Romania was chosen as case study area. Landsat scenes from the year 2013 are grouped based on vegetation phenology into four PGSs (periods of the growing season). NDVI values corresponding to land abandonment are analyzed with Classification and Regression Trees. A total of 23 models—representing combinations of PGSs—are tested in order to determine what period of the vegetation growing season fits best for mapping abandoned land, and for the purpose of deriving such a map for Bucharest. Finally, results are validated against independent data and a resource estimation for the entire mapping process is performed. Results show that abandoned land can be mapped with Landsat imagery with accuracies above 80%. Higher accuracies are obtained when scenes encompassing the beginning of the vegetation growing season are included in the models. We also observed that accuracy tends to decrease from models which include PGSs representing the beginning of the vegetation growing season towards those representing the end. Estimation showed that mapping land abandonment with Landsat data could reduce time and workforce resources by almost half compared with aerial imagery and field work. As our method is rapid, easy to implement, and based on freely available data, it can be used by local authorities that cannot allocate significant resources for land change monitoring. Furthermore, the approach could provide objective information to municipalities where official statistics on land abandonment are unreliable or of low quality. [ABSTRACT FROM AUTHOR]

• Trend analysis confirmed the greening trend of vegetation at spatio-temporal. • Bioclimatic variables revealed the effects of climate more clearly. • Bio12, LUCC, DEM and Pd are the more important variables of vegetation greening. • The BRT revealed specific nonlinear response thresholds for diverse driving factors. The main influencing factors and their nonlinear effects on the changes of vegetation in China's mountainous areas under the interaction of different factors are not yet clear, and comprehending the evolutionary trends and driving mechanisms of vegetation is crucial to reveal the changes in ecosystem structure and function. In this study, trend analysis (M−K, T-S and EEMD) combined with machine learning algorithm, namely Boosted Regression Tree model (BRT), were used to quantify the trends of nonlinear responses and thresholds for bioclimatic variables, topography, soil properties and anthropogenic factors for vegetation changes in Longnan. The results showed that the trend analysis clearly confirm the increasing trend of vegetation at multiple spatio-temporal scales. The BRT indicated that total precipitation (bio12, 15.22%), land use (LUCC, 12.68%), elevation (DEM, 11.20%), and population density (Pd, 9.20%) were the more important factors of dominant vegetation greening. Bioclimatic variables were found to revealed the effects of climate with vegetation more clearly. In addition, the BRT revealed that the selected factors have the different nonlinear response relationships to vegetation greening trend and specific thresholds. Among them, increasing of cropland, grassland and forestland can promote vegetation greening. However, GDP, Pd, DEM, bio12, mean diurnal range and temperature seasonality (bio2, bio4) exceed the threshold can significantly inhibit vegetation growth. The BRT combined with trend analysis revealed the nonlinear response relationships and thresholds of the drivers behind the vegetation change patterns, which have obvious effects in exploring the driving mechanisms of vegetation changes in mountainous areas. This study provided an important reference for better revealing the interaction mechanisms between vegetation changes and drivers in the semi-humid zone in East Asia even globally. [ABSTRACT FROM AUTHOR]

Soil organic carbon (SOC) is pivotal for biological, chemical and physical processes and provides vital information on changes in soil fertility and land degradation. Rangelands, accounting for about 81% of Australian land area, are significant carbon (C) stores and small increases in soil C sequestration over such a vast area represents a considerable climate change mitigation opportunity. Efficient modelling techniques to evaluate the potential to increase rangeland SOC stocks are vitally important to assess their role in the global carbon cycle and quantum abatement. This study aimed to evaluate boosted regression trees (BRT) and random forest (RF) models in predicting SOC stocks from available continuous remotely sensed variables using two feature selection techniques. Dominant variables that affect SOC stocks in the rangelands were also identified. Using field-based measurements of SOC stock collected from 564 data points across the study area and 28 GIS-based environmental variables including climate, topography, radiometry, vegetation and land fractional cover data, we employed stepwise regression (SR, linear approach) and genetic algorithm (GA, nonlinear approach) to select the most informative variables. These selected predictors were then used to train the BRT and RF models. In all, four models were evaluated: BRT using SR (SR_BRT); RF using SR (SR_RF); BRT using GA (GA_BRT) and RF using GA (GA_RF). In addition, BRT using all predictors (All_BRT) and the RF using all predictors (All_RF) were used as benchmarks to test the performance of the four models. Of the field-based data, 75% were used to train the model (“calibration dataset”) and the remaining 25% were used to validate the prediction of SOC stocks (“validation dataset”). The results indicate that the RF exhibited a better performance in predicting SOC stocks than the BRT regardless of input variables. In addition, we verified that feature selection for both machine learning techniques is necessary for estimating SOC stocks because they can increase accuracy and save time. The GA_RF was the most reliable method to predict SOC stocks, with the lowest root mean square error (RMSE) and the highest R 2 values (7.44 Mg C ha −1 and 0.48, respectively), suggesting that the method of using GA-RF to generate a predictive model from measured data and remotely-sensed variables may provide a cost effective alternative to direct sampling to predict SOC stocks in the semi-arid rangelands of eastern Australia. The important variables for explaining the observed SOC stocks were rainfall, elevation, Prescott index (PI, a measure of water balance), and land fractional cover (bare ground fraction). The approach proposed here can be extended in areas where field observed data is scarce (e.g. rangelands) to produce more detailed information about SOC stocks. As such, the results of our study are of particular importance in Australian rangelands to provide a statistical and theoretical basis for producing digital SOC stock maps based on readily available remotely-sensed data, with potential for use in similar rangelands conditions internationally. [ABSTRACT FROM AUTHOR]

Water resource managers face increasing challenges in identifying what physical and chemical stressors are responsible for the alteration of biological conditions in streams. The objective of this study was to assess the comparative influence of multiple stressors on benthic diatoms at 98 sites that spanned a range of stressors in an agriculturally dominated region in the upper Midwest, USA. The primary stressors of interest included: nutrients, herbicides and fungicides, sediment, and streamflow; although the influence of physical habitat was incorporated in the assessment. Boosted Regression Tree was used to examine both the sensitivity of various diatom metrics and the relative importance of the primary stressors. Percent Sensitive Taxa, percent Highly Motile Taxa, and percent High Phosphorus Taxa had the strongest response to stressors. Habitat and total phosphorous were the most common discriminators of diatom metrics, with herbicides as secondary factors. A Classification and Regression Tree (CART) model was used to examine conditional relations among stressors and indicated that fine-grain streams had a lower percentage of Sensitive Taxa than coarse-grain streams, with Sensitive Taxa decreasing further with increased water temperature (>30 °C) and triazine concentrations (>1500 ng/L). In contrast, streams dominated by coarse-grain substrate contained a higher percentage of Sensitive Taxa, with relative abundance increasing with lower water temperatures (<29 °C) and shallower water depth (<0.3 m). Quantile regression indicated that maximum water temperature appears to be a major limiting factor in Midwest streams; whereas both total phosphorus and percent fines showed a slight subsidy-stress response. While using benthic algae for assessing stream quality can be challenging, field-based studies can elucidate stressor effects and interactions when the response variables are appropriate, sufficient stressor resolution is achieved, and the number and type of sites represent a gradient of stressor conditions and at least a quasi-factorial design. [ABSTRACT FROM AUTHOR]

[Display omitted] • EVI × PAR shows great potential for tracking the GPP variations in coastal wetlands based on the simple linear regression. • NDVI × PAR overtakes EVI × PAR for the GPP estimation in the salt marshes with little tidal influence and plant litter. • Random forest regression improves the performance in upscaling the GPP measurements to large spatial scales. • High spatial resolution, a good combination of VIs, and advanced algorithms are vital for estimating GPP in coastal wetlands. Accurate monitoring of the spatiotemporal variations in the gross primary productivity (GPP) in coastal wetlands is essential for blue carbon quantification. However, the currently available moderate-spatial-resolution GPP algorithms and products may contain large uncertainties and may rarely meet the demands of tidal wetland research. In this paper, we examine two statistical methods, namely, simple linear regression and random forest regression, to capture the high spatial resolution GPP variations in the salt marshes in the Yangtze Estuary by integrating multi-source data, including eddy covariance, Landsat, and Moderate-resolution Imaging Spectroradiometer (MODIS) satellite data. The satellite-derived photosynthetically active radiation (PAR), enhanced vegetation index (EVI), normalized difference vegetation index (NDVI), and normalized difference water index (NDWI) are used individually and in different combinations to drive the two statistical methods and investigate their performances in estimating the GPP. The results show that together with the PAR, the EVI generally has the greatest potential for GPP estimation (R2 > 0.75, RMSE < 6.80 μmol m−2 s−1) based on simple linear regression. However, the NDVI outperforms the EVI in wetlands with little tidal flooding and plant litter. The random forest method improves the performance in terms of upscaling the GPP measurements to large regions, and the combination of the EVI, NDVI, NDWI, and PAR performs the best, indicating that the use of the more advanced algorithm and the incorporation of a remotely sensed index that reflects the tidal influence are beneficial to GPP estimation in coastal salt marshes. Moreover, using Landsat data with a high spatial resolution (30 m) yields a much more accurate GPP estimation than using MODIS data with a 500 m resolution. Our results demonstrate that the use of high spatial resolution data, the proper use of remotely sensed indices, and the incorporation of a good combination of these indices through an advanced algorithm such as a machine learning algorithm are vital for capturing the nearly instantaneous GPP variations in coastal wetlands at large spatial scales. [ABSTRACT FROM AUTHOR]

[Display omitted] • Predicting SOC and SIC from two sampling periods (1990 vs 2015); • Quantifying the impact of environmental variables on SOC and SIC changes; • Mapping soil carbon stocks change during 25 years (1990–2015). Soil organic carbon (SOC) and soil inorganic carbon (SIC) has important effects on soil physical, chemical and biological properties. They play an important role in coordinating the relationship between soil water and air, increasing soil water holding capacity and improving plant productivity. In this study, a boosted regression trees (BRT) model was developed to map the spatial distribution carbon stocks in the semi-arid region of Northeast China in 1990 and 2015. During the two periods, 10-fold cross-validation technology was used to test the performance and uncertainty of BRT model. In order to construct the model, 9 environmental variables (derived from climate, topography and biology) and 173 (1990) and 223 (2015) topsoil (0–30 cm) samples were used. The comparison between estimated and observed data shows that the RMSE of SOC and SIC stocks were 0.53 kgm−2 and 0.19 kgm−2 in 1990, and 0.65 kgm−2 and 0.20 kgm−2 in 2015, respectively. Elevation, normalized difference vegetation index, mean annual precipitation and Landsat band 3 were identifies as critical environmental factors for simulating the spatial distribution of SOC, accounting for 76.6 % and 70.3 % of the total relative importance in 1990 and 2015, respectively. Mean annual precipitation, mean annual temperature and topographic wetness index were the critical environmental factors for simulating the spatial variation of SIC during the two periods. Land use change also played an important role in the spatial variability of SOC and SIC stocks. In the past 25 years, soil carbon stocks decreased from 6.2 kg m−2 in 1990 to 5.9 kg m−2 in 2015. The spatial distribution pattern of SOC was high in northeastern area and low in southwestern area during the two periods, while the spatial distribution pattern of SIC was opposite to that of SOC stocks. The mapped soil carbon stock distribution is fundamental to future study of soil carbon cycle and regional carbon balance in semi-arid regions. [ABSTRACT FROM AUTHOR]

[Display omitted] • A hybrid CART-GAMs model is proposed to evaluate habitat suitability. • CART can identify critical habitat factors to reduce data redundancy. • GAMs can quantify the nonlinear response of species to habitat factors. • This model was applied to the Pearl River Estuary to confirm its validity. Benthic macroinvertebrates are an important part of the aquatic food chain in estuaries. Therefore, understanding their associated habitat requirements is critical in maintaining estuarine ecosystem stability, which has until now been evaluated using habitat suitability evaluations. Although most habitat suitability index (HSI)-based methods can determine the target species preferences for a few specific habitat factors, benthic macroinvertebrates are influenced by multiple habitat factors in estuaries. However, applying too many habitat factors can result in data redundancy, thereby reducing model efficiency. Additionally, nonlinear relationships between species and habitat factors and interactions between different habitat factors are often difficult to evaluate. Previous studies have not focused on methods that can simultaneously address these two issues. Accordingly, this study introduced CART-GAMs, a classification and regression tree model (CART) and a generalized additive model (GAM) hybrid, which can evaluate the suitability of multiple habitat factors relevant to benthic macroinvertebrates in estuaries. First, to remove irrelevant variables (i.e., data redundancy), CART was used to identify critical habitat factors that have significant impacts on benthic macroinvertebrates. GAMs were then used to quantify the nonlinear relationships between critical habitat factors and species. Finally, response curves fitted by GAMs were used to determine species habitat preferences (i.e., suitable critical habitat factor ranges). This study applied the CART-GAMs model to the Pearl River Estuary (PRE) to test its effectiveness. The results showed that CART identified four critical water quality factors, that is, salinity (SAL), total phosphorus (TP), pH, and ammonia nitrogen (NH 3 -N) from the 12 measured factors, subsequently reducing data redundancy by 66.7%. Using critical water quality factors, GAMs explained 66.2% of the variance in Margalef's diversity index (dM) of benthic macroinvertebrates. Additionally, the GAM response curves showed that dM was positively correlated with SAL and negatively correlated with TP and TN. The response curve between dM and pH is unimodal. Finally, this study determined suitable water quality factor ranges for benthic macroinvertebrates (SAL = 21.15‰ ∼ 32.32‰, TP = 0.05 ∼ 0.2 mg/L, pH = 7.7 ∼ 8.2, and TN = 0.4 ∼ 1.5 mg/L). This study demonstrates that the hybrid CART-GAMs model can effectively evaluate estuarine habitat suitability with multiple habitat factors, while providing a reference for estuary habitat management. [ABSTRACT FROM AUTHOR]

Forests are living dynamic systems and these unique ecosystems are essential for life on earth. Forest fires are one of the major environmental concerns, economic, and social in the worldwide. The aim of current research is to identify general indicators influencing on forest fire and compare forest fire susceptibility maps based on the boosted regression tree (BRT), generalized additive model (GAM), and random forest (RF) data mining models in the Minudasht Township, Golestan Province, Iran. According to expert opinion and literature review, fifteen condition factors on forest fire have been selected in the study area. These are slope degree, slope aspect, elevation, topographic wetness index (TWI), topographic position index (TPI), plan curvature, wind effect, annual temperature and rainfall, soil texture, distance to roads, rivers, and villages, normalized difference vegetation index (NDVI), and land use. Forest fire locations were identified using MODIS images, historical records, and extensive field checking. 106 (≈70%) locations, out of 151 forest fires identified, were used for models building/training, while the remaining 45 (≈30%) cases were used for the models validation. BRT, GAM, and RF data mining models were used to distinguish between presence and absence of forest fires and its mapping. These algorithms were used to perform feature selection in order to reveal the variables that contribute more to forest fire occurrence. Finally, for validation of models, the area under the curve (AUC) for forest fire susceptibility maps was calculated. The validation of results showed that AUC for three mentioned models varies from 0.7279 to 0.8770 (AUC BRT = 80.84%, AUC GAM = 87.70%, and AUC RF = 72.79%,). Results indicated that the main drivers of forest fire occurrence were annual rainfall, distance to roads, and land use factors. The results can be applied to primary warning, fire suppression resource planning, and allocation work. [ABSTRACT FROM AUTHOR]

Nonparametric approaches including a classification and regression tree (CART), a nonparametric changepoint analysis (nCPA) and a Bayesian hierarchical modeling (BHM) method were developed to determine ecoregional nutrient response thresholds. A CART analysis revealed that hierarchical structure was important for predicting Chl a concentrations from total nitrogen (TN) and total phosphorus (TP). The nCPA and BHM methods confirmed the CART results for each node in the tree, and the 90% confidence interval for each threshold was calculated to quantify uncertainty. The CART, nCPA, and BHM methods suggested that the nutrient criteria differed significantly within certain nutrient ecoregions and that numerical nutrient criteria of 0.0150–0.222 mg/L TP and 0.300–1.766 mg/L TN may control Chl a concentrations in the various lake ecoregions. The results of this analysis suggest that the integration of CART, nCPA and BHM might be useful for determining nutrient thresholds. [ABSTRACT FROM AUTHOR]

• Boosted regression tree model is robust in charactering nonlinear non-point source pollution generation and transport processes. • The nonlinear responses of critical source areas to influencing factors was evaluated quantitatively. • The thresholds for influencing factors can provide supportive information for watershed management. • Land use and fertilizer application have higher importance in determining the occurrence of critical source areas. • Machine learning techniques have great potential for predicting critical source areas under climate change. Critical Source Areas (CSAs) are areas that contribute disproportionate high levels of non-point source (NPS) pollution to receiving waters, and their occurrence is the result of the complex interaction between the factors related to the sources and transport processes of NPS pollution. A systematic understanding of how these influencing factors affect CSAs is essential for successful watershed management. In this study, we applied a statistical data mining technique boosted regression tree model to quantify the contribution of eight influencing factors (soil type, slope, elevation, RUSLE LS factor, RUSLE K factor, runoff, fertilizer application rate and land use) on two types of CSAs (TN-CSAs and TP-CSAs), as well as the marginal effects and potential thresholds of influencing factors on the occurrence of CSAs. Results show that land use (37.35%, 25.03%), fertilizer application (36.93%, 57.83%) and soil type (17.59%, 13.70%) have higher importance in determining the occurrence of TN-CSAs and TP-CSAs; and the incidence of TN-CSAs is positively correlated with most factors before the threshold for each influencing factor, after which the marginal effect largely leveled off or dropped slightly; TP-CSAs have essentially the same characteristics as TN-CSAs, but TP-CSAs are more likely to occur in areas with an annual runoff of around 244.92 mm. In addition, this study discussed the application of machine learning techniques in predicting CSAs under climate change without physical-based models, as well as a preliminary watershed management planning for NPS pollution control in the study watershed. These results provided important information for nutrient management regulations. [ABSTRACT FROM AUTHOR]

• Vegetation coverage of desert ecosystems was investigated in the Qinghai-Tibet Plateau (QTP). • A boosted regression tree (BRT) model was employed for desert vegetation identification. • A robust empirical model based on in-situ FVC and satellite-derived MSAVI was proposed. • Currently available vegetation coverage estimates for QTP's desert ecosystems underestimate FVC by at least 29%. Accurate fractional vegetation coverage (FVC) detection is beneficial for evaluating the dynamics and development potential of desert ecosystem. However, at present the performance of the related published products in the Qinghai-Tibet Plateau (QTP)'s desert ecosystem has not been verified, and the method for extracting FVC for this region from unmanned aerial vehicle (UAV) RGB images has also not been tested. UAV RGB images were processed via a boosted regression tree model (BRT). 57 ground measurements were collected, and were further paired with vegetation indexes (VIs) from satellite sensors to develop an empirical model that was used to evaluate FVC of QTP's desert ecosystem. The results showed that: (1) BRT effectively enhanced the vegetation information in UAV RGB images and improved the accuracy of FVC ground measurements (the area under the receiver operating characteristic curve (AUC) = 0.95, kappa > 0.95); (2) The relationship between FVC and modified soil-adjusted vegetation index (MSAVI) was most robust in QTP's desert ecosystem, and a power model (FVC = 13.85 × MSAVI2.07, R2 = 0.89, RMSE < 0.01) was proposed to derive local FVC within an area of 1.05 × 106 km2; (3) Although the current published FVC products show spatial consistency in QTP's desert ecosystem, FVC was underestimated by these products by at least 29% compared with the FVC values derived from the power model determined in this study. [ABSTRACT FROM AUTHOR]

• Ecological Potential (EP) data are needed to serve as a baseline for comparison. • EP reflects potential cover under least disturbed and most productive conditions. • EP models were trained on intact sites with no recent disturbance or treatment. • Departure of contemporary cover from EP can be used as a degradation screening tool. • EP data can provide monitoring and control to evaluate treatment effectiveness. Rangelands have immense inherent spatial and temporal variability, yet land condition and trends are often assessed at a limited number of spatially "representative" points. Spatially comprehensive, and quantitative, Ecological Potential (EP) data provide a baseline for comparison to current rangeland vegetation conditions and trends. Here, we define EP as potential fractional cover (bare ground, herbaceous, litter, shrub, and sagebrush) represented in the least disturbed areas and most productive years of the Landsat satellite archive (1985-present) for each 30-m pixel. We produce EP maps across rangelands in the western United States by training regression tree models using Rangeland Condition Monitoring Assessment and Projection (RCMAP) time-series fractional cover maps in ecologically intact sites (with limited annual herbaceous cover, no recent disturbance or vegetation treatment, and less bare ground cover than expected). As independent predictor variables in these models, we use digital soils and topography data and six bimonthly composites of the 90th percentile of Normalized Difference Vegetation Index (NDVI) and associated spectral bands from the 1985–2020 Landsat archive. EP predictions were successful in capturing biophysical gradients present in the independent variables and depicting potential cover in the absence of disturbance; we found no influence of fires or land treatments in the data. Next, we compared EP to contemporary (2018) cover, to create departure maps that can be used as a screening tool indicating degradation and providing an early warning of vegetation state change. Finally, we used a dichotomous key to convert the 1985 and 2018 RCMAP cover and EP cover into vegetation states important to land management decisions (invaded sagebrush steppe, annual grasslands, etc.). We found that in 1985, 21.2% of the study area had a different vegetation state than EP, and this percentage increased to 24.2% by 2018. More than 50% of the EP native sagebrush steppe was converted to an annual grassland, perennial grassland, or non-sagebrush shrub by 2018, and an additional 7% was classified as invaded sagebrush steppe, at risk of transition to another state. EP products provide a spatio-temporal reference of vegetation conditions from the last three decades across rangelands in the western United States. Use of the EP reference can improve adaptive management practice by providing monitoring and control data, which are often lacking, and assist in differentiating treatment effect from confounding factors. [ABSTRACT FROM AUTHOR]

[Display omitted] • The determinants of MHT were identified both statistically and pedologically. • Distribution patterns of MHT in alpine environment were first revealed. • Partial dependence plots revealed the non-linear predictor-response relationships. • The proposed approach can be adopted into the Digital Soil Mapping framework. Mattic horizon, grassroots intertwined surface horizon for soils in alpine meadow, is of great importance for alpine ecosystems with its thickness serving as a critical indicator for the extent of alpine meadow degradation. However, the distribution patterns and key environmental factors controlling mattic horizon thickness (MHT) remain unknown. We evaluated the potential of a boosted regression trees (BRT) based variable selection approach to identify the determinants of MHT, to reveal the predictor-response relationship, and to map the spatial distributions of MHT. Results showed that the selected six key environmental variables can reliably reflect the pedogenic processes of mattic horizon in such alpine environment. Among which, the most important environmental variables controlling MHT were modified soil-adjusted vegetation index 2 (relative importance of 31.9%), elevation (22.7%), followed by slope (13.2%), topographic wetness index (11.4%), relative slope position (10.4%), and mean annual precipitation (10.4%). Additionally, the BRT model using those six key variables for mapping MHT performed an acceptable prediction with R2 value of 0.447 (standard deviation of 0.019). The variable selection procedure dramatically reduced the number of environmental variables by 77% (from 26 to 6), which could not only improve computational efficiency but also significantly improve model performance with a 30.3% increase in R2. Our results demonstrate that the proposed BRT algorithm-based variable selection approach can reliably identify the determinants of MHT both statistically and pedologically. It can be used as a promising approach for revealing the predictor-response relationship, mapping the distribution patterns of MHT, and may provide a solution to map soil depth over large areas. [ABSTRACT FROM AUTHOR]

[Display omitted] • Current speed and wave orbital velocity are important indicators of fish abundance. • A hectare of seagrass produces 110 – 1,080 kg−1 of fish per year. • A hectare of seagrass supports 69 – 865 fishing trips per year. • Studied 6662 ha of seagrass are valued at AUD 36,782 million annually. The value of critical habitats, such as seagrass, to act as a nursery varies spatially and temporally; however, such information is essential for the public and stakeholders to appropriately value and manage these habitats. We use an existing systematic long-term fisheries dataset in Port Phillip Bay to examine variability in nursery habitat value for an important commercial and recreational species, King George Whiting (Sillaginodes punctatus). Port Phillip Bay represents one of the most important marine assets in the southern hemisphere and is surrounded by the second-largest city in Australia, Melbourne, home to 4.5 million people. We modelled the abundance of King George whiting as a function of environmental variables, using Boosted Regression Trees (BRT). Fish densities ranged from 1,000 to 30,000 individuals ha−1y−1, equalling an adult biomass of 110–3,300 kg ha−1y−1. This production supports between 69 and 2,062 recreational fishing trips a year, with an estimated value of seagrass of AUD 687–20,625 ha−1y−1. Based on biomass production of King George Whiting and recreational fisheries data, the 6662 ha of seagrass in Port Phillip Bay are valued at around AUD 36 million annually. [ABSTRACT FROM AUTHOR]

We used data from regional forest inventories and research programs, coupled with mapped climatic and topographic information, to explore relationships and develop multiple linear regression (MLR) and regression tree models for total and deciduous shrub cover in the Oregon coastal province. Results from both types of models indicate that forest structure variables were most important for explaining both total and deciduous shrub cover. Four relationships were noted: (1) shrub cover was negatively associated with Tsuga heterophylla basal area and density of shade-tolerant trees; (2) shrub cover was negatively associated with variables that characteristically peak during stem exclusion and mid-succession; (3) shrub cover was positively associated with variables that characterize later successional stages; and (4) higher total and especially deciduous shrub cover were positively associated with hardwood stands. Environmental variables were more important for explaining deciduous shrub cover compared to total shrub cover, but they have an indirect effect on total shrub cover by influencing tree composition. However, because of land ownership patterns, it was difficult to decouple environmental from disturbance factors associated with management strategies across multiple ownerships.Tree models performed similarly (PRD=0.17–0.27) or better compared to MLR models (PRD=0.17–0.23) although they contained more (2) predictor variables. Our results indicate that response variable transformation can greatly improve regression tree model performance. While interpretation of MLR and tree models were somewhat similar, the tree models allowed a more explicit understanding of relationships and provided thresholds for anticipating shifts in shrub cover. Such thresholds are useful to forest managers who are monitoring and evaluating critical amounts of shrub cover necessary for different ecosystem components such as bird habitat. Lack of strong predictive power in both types of models may be because many common shrubs can persist and maintain consistent cover under a variety of stand and environmental conditions or there may be a lag time between disturbance events and shrub response. The stochastic nature of disturbance and their interactions with site conditions also makes prediction at this scale in this highly managed landscape inherently problematic. Yet, our models provide both a predictive and conceptual tool for understanding shrub cover patterns across the region. [Copyright &y& Elsevier]

• Surface Urban Heat Island (SUHI) are calculated through land surface temperature. • Potential SUHI influencing factors are explored in a multidirectional perspective. • Contribution rates of influencing factors to SUHI in the 16 cities is identified. • Multiple influencing factors have linear or nonlinear interactions among them. Nowadays urban climate is a global problem and many studies focused on understanding the relation between urban climate the built-up space using radiometric observations of the land surface temperature to estimate and monitor the surface urban heat island intensity (SUHIs). In this study MODIS land surface temperature (LST) data were used. The Yangtze River Delta Urban Agglomeration (YRDUA), eastern China, was selected as an example to study SUHI and multiple influencing factors in 16 big cities. Anthropogenic factors are considered the most important ones in determining SUHI, while natural factors remain influential. By using stratified random sampling (SRS), 78,085 random points were selected within the 16 cities. Nine influencing factors were selected in this study: distance from building (BD), distance from the main roads (RD), distance from water (WD), digital elevation model product (DEM), gross domestic product (GDP), normalized difference vegetation index product (NDVI), nighttime lighting intensity (NTI), population (POP) and impervious surface area data (%ISA). The SUHI intensity was extracted at each random point as well as the values of the influencing factors, NDVI, DEM, ISA, POP, NTI and GDP. For BD, WD and RD, random points were selected from the water, building and main roads using the near tool in ArcGIS to measure these distances. Boosted regression tree (BRT) model was applied to capture the contributions of the above factors to SUHI. We also applied a different procedure to evaluate the relative influence of Land Use and Land Cover (LULC). The relative influence refers to the contribution of each factor to determine SUHI. The influencing factors were ranked on the basis of the relative influence on SUHI. The results showed that (1) higher SUHI intensity was recorded in Shanghai, Jiaxing and Nanjing cities respectively, while Hangzhou recorded the lowest SUHI. (2) Anthropogenic drivers have slightly higher relative influence on SUHI than natural drivers, i.e. 51.29% and 48.71% respectively. The influence of all drivers on SUHI from high to low is NTI (27.62%), ISA (24.38%), NDVI (12.11%), GDP (7.95%), DEM (7.29%), POP (6.37%), BD (5.33%), WD (4.93%), RD (4.02%). (3) The variation in the socioeconomic level lead to different spatial patterns of different influence factors, further indicating that the overall mean SUHI intensity is affected by the development of the city. [ABSTRACT FROM AUTHOR]

• Spatio-temporal patterns in biodiversity were mapped in a peri -urban environment. • Ecoacoustic index data were interpolated and modelled using boosted regression trees. • Landform, vegetation type and extent were key predictors of biodiversity. • Ecoacoustic analyses can guide conservation decision-making and urban planning. As the rate of urbanisation continues to increase, widespread habitat clearing within peri-urban landscapes contributes to significant environmental impacts, including loss of biological diversity. Acoustic recording has recently been identified as an effective tool for monitoring biodiversity and ecosystem health. With increasing pressure from urbanisation, it is critical that spatial and temporal variability in biodiversity is mapped across future development sites to enable sound decision-making and to deliver ecological urban design outcomes. This study used ecoacoustic monitoring to map biodiversity patterns in space and time to identify hot spots and hot moments of biodiversity activity across a peri-urban landscape in south-east Queensland, Australia. In this study, a hot spot represents an increase in acoustic activity at a given spatial location, whereas hot moments represent an increase in acoustic activity at a given time point. An acoustic index (Acoustic Complexity Index, ACI) was used as a proxy for biodiversity and visualised through spatial interpolation. The acoustic data were statistically modelled using Boosted Regression Trees (BRT). This approach enabled predictors related to acoustic complexity to be identified, including vegetation and landform. Results of this study have shown that ecoacoustic data can be used to map hot spots and hot moments of biodiversity and support more informed conservation decision-making in future urban planning frameworks, to avoid or mitigate negative impacts on biodiversity. [ABSTRACT FROM AUTHOR]

• A classification approach for urban metabolic types was presented in this study. • GDP per capita and climate types were important threshold variables for classifying urban resource consumption levels. • Different types of the features of urban metabolism were usually associated with specific urban resources. • The potential influence of climatic conditions should be considered in urban resource utilization strategies. Exploring resource utilization strategies for different types of cities is essential for achieving the sustainable development of cities. Urban metabolism is a systems-oriented approach to understanding the interrelationship between urban socioeconomic status and natural environment through resource utilization. Recognition and identification of urban metabolic types could help in formulating targeted urban resource utilization strategies. However, few studies have investigated the classification of urban metabolism. Therefore, this study combines material flow analysis with the classification and regression tree model to present a classification approach for urban metabolism that considers cities' socioeconomic, demographic, and climatic conditions and resource consumption. This approach was tested in Chinese prefecture-level cities. The results showed that gross domestic product (GDP) per capita, population, population density, and climate types were significantly correlated with cities' resource consumption indicators. GDP per capita and climate types were the most important threshold variables for classifying urban resource consumption levels. GDP per capita was the optimal split node for electricity, total energy, water, industrial minerals and metals, and total outputs, while climate types was the optimal split node for biomass, fossil fuels, and construction materials. The level of urban resource consumption in temperate climates was higher than in other climate types. Eight types of urban metabolism were identified in China that were closely related to urban socioeconomic conditions, scale grade, resource endowment, industrial level, and energy structure. The findings of this study provide a reference for determining the problem of urban critical resource consumption and formulating appropriate urban resource utilization strategies. [ABSTRACT FROM AUTHOR]

• Health of Trichocereus atacamensis specimens is assessed by a 3-step decision tree. • Local citizen knowledge plays a critical role in evaluation for conservation policies. • Multiple Regression Trees analyses improve health assessment protocols for cacti. • Health situation of T. atacamensis is multifactorial and needs special attention. Argentina is considered megadiverse for Cactaceae, and Jujuy province, with a high proportion of cacti endemism, is a critical region for their preservation. Cacti deterioration is mainly associated with habitat degradation, agricultural frontier advance, urbanization, illegal collection and trade. Of the three species of columnar cacti within Trichocereus genus that are present, Trichocereus atacamensis represents a valuable species for local communities. In the last decades, detrimental effects on T. atacamensis populations increased despite the reduction in the use of timber for construction or handicrafts. Knowledge on the health of cacti populations is fundamental for conservation and resource management, therefore the objective of the present study was to develop a health status classification criterion for cacti in Quebrada de Humahuaca, Jujuy. A generic method was elaborated to assess the health of cacti populations to facilitate and speed-up the decision-making process for conservation purposes. The evaluation of factors involved in the deterioration is very complex. We designed a method considering aspects that the local residents identified as the most relevant: a) presence of air plant (genus Tillandsia), b) larvae of Cactoblastis bucyrus and c) signs of rot. Band transects were sampled in patches of T. atacamensis , surveying these aspects for each individual. A Multivariate Regression Tree analysis (MRT) was conducted. According to this clustering, an accessible assessment protocol with a 5-level health status was established to classify the cactus as optimal, suboptimal, compromised, vulnerable or critical. The general health status of the patches was highly variable. This MRT analysis is potentially applicable to any columnar cacti population and to establish the basis of a decision protocol for health status assessment. It is evident that the situation of T. atacamensis transcends as a local problem and exceeds the presence of pathogens suggesting a multifactorial causality that needs special attention. [ABSTRACT FROM AUTHOR]

• Functional traits described better human intervention than species identity. • Flow-regulation by river damming increased the functional nestedness of phytoplankton. • Species from dammed areas were filtered by the river flow after the dams. • Land use amplified the effects of river damming on the phytoplankton community. River damming reduces the dispersal rates of biota and filter species by altering the local environmental conditions. Although phytoplankton is a key biotic group to indicate changes in aquatic environments, the effects of river damming on the longitudinal distribution of phytoplankton communities are not widely addressed. Here we investigated the changes on the taxonomic and functional composition of the phytoplankton community in a large tropical dammed river. We expect compositional changes to be higher at taxonomic than functional level especially in the dam-free stretches of the river because running waters select species able to cope with the water turbulence. Furthermore, flow regulation allows species with different traits to colonize the increasingly lentic stretches resulting in higher functional diversity towards dams. Phytoplankton samples were taken in 210 sites, spaced by 5 km, along the whole river channel. Data were split at a reservoir level considering the sites within 30 km before and after each dam to analyze the effects of damming on the community composition. Compositional changes were assessed by estimating the degree to which each community differed from all other communities within the regional pool (community distinctiveness) considering taxonomic and functional composition separately. Functional diversity was estimated as the degree to which species within the same community differed from each other with respect to their trait composition (functional dispersion). We used community-weighted means to test how the different traits were filtered by river damming. Furthermore, we assessed the effects of local environmental conditions on the phytoplankton taxonomic and functional composition using boosted regression trees (BRT). We observed no effects of river damming on species dispersal. Functional turnover was low, suggesting that compositional changes occurred mainly among functionally similar species. However, there was a noticeable reduction of phytoplankton diversity that persisted for more than 100 km after one of the dams. Flow regulation allowed flagellated and mucilaginous species to colonize the dammed areas. After the dams, such species were filtered due to the increase of water mixing downstream, as evidenced by a hump-shaped trend in the functional diversity of the sites within 30 km distance of the dams. BRT analyses highlighted the role of nutrient input in the dammed areas for the community compositional change. This study highlights the importance of considering both taxonomical and functional diversity to describe the effects of human intervention such as damming on aquatic ecosystems. [ABSTRACT FROM AUTHOR]

• We test approaches for tree selection in allometric biomass models. • Sample size and size-distribution were found to be important for prediction accuracy. • Uniformly distributed sample trees on D-range were optimal for model development. • The extent and position of D-range affected the goodness of fit of models. • R2 was a poor indicator of prediction performance for allometric models. Developing allometric biomass models is an important process because reliability of forest biomass and carbon estimations largely depend on the accuracy and precision of such models. The effects of tree sampling on tree aboveground biomass (AGB) prediction accuracy and precision are complex and can, therefore, be difficult to quantify. In this paper we use a Monte Carlo simulation to investigate how model prediction accuracy and precision are affected by tree sampling approaches. Because diameter at breast height (D, in cm) is the most common predictor of tree AGB (in kg dry weight), we focused our analysis on the AGB-D relationship. The following sample characteristics were investigated: (i) sample size; (ii) extent of the D-range (difference between the largest and the smallest D value); (iii) position of D-range (characterized by the starting point of D-range); and (iv) the size-distribution (distribution of D) of sample trees. We found that, although the natural variability of AGB-D relationship was a key driver for both prediction accuracy and precision, the above sample characteristics were important for improving prediction accuracy. Although having a negligible effect on precision, both sample size and size-distribution of sample trees, greatly influenced prediction accuracy. We demonstrate that selecting a constant number of trees for each D class (i.e. uniform distribution of the sample trees over the D-range) generally produced models that were more accurate predictors of AGB. The extent and position of D-range, although considerably affecting the goodness of fit and the standard errors of allometric model parameters, had only a marginal effect on AGB prediction accuracy and precision. Furthermore, we showed that R2 was a poor indicator of model prediction accuracy and precision, due to its sensitivity to changes in D-range. These findings inform certain practical recommendations we report for improving the accuracy and precision of biomass prediction. [ABSTRACT FROM AUTHOR]

A workflow of the research method process. • Bioclimatic variables are very suitable for examining climatic impacts. • It is needed to differentiate human impacts of different kinds. • Importance of climate change, human activity and topography was 42.7%, 33.2%, 18.5%. • All nature and human factors showed nonlinear relationship with vegetation greening. • Future climate changes will facilitate vegetation greening. Vegetation showed a greening trend in most Southwest China. However, the nonlinear relationship of vegetation greening with nature and human factors remains unclear. In this study, we studied the nonlinear relationship and predicted the future changes of the greening with Boosted Regression Tree (BRT) method. Results showed that: (1) Precipitation of Driest Month (Bio14, 20.64%), land use changes (10.39%) and population density (8%) were the three most important factors limiting vegetation greening. Climate changes (42.655%) and human activities (33.163%) were the two most important variable types, but topography was also important (18.481%), which cannot be ignored; (2) Bio14 and elevation had the strongest variable interactions. Climate changes had strong interactions with both human activities and elevation, but human activities and elevation had less interactions; (3) vegetation greening was facilitated by the increasing of Bio14, distance to residential area, temperature annual range (Bio7), but inhibited by the increasing of GDP and precipitation of coldest quarter (Bio19) with changing rates. These factors would have no further impacts when approaching threshold values. The increase of population density improved vegetation greening greatly when it was low, while inhibited the greening strongly when it was high. Elevation increase promoted vegetation greening when elevation <300 m, then inhibited it. For land use changes, 'Grain for Green' improved vegetation greening, but urbanization decreased it; (4) Under current environmental condition, area percentage of vegetation browning will greatly increase from 13.88% to 37.69%, which is mostly from insignificant vegetation changes and located in the west. However, future climate changes will facilitate vegetation greening in the most area except the northwest. Our results highlight the importance of nonlinear analysis for determining the drivers and predicting future changes of vegetation greening, and developing adaptation and alleviation strategies for climate changes and human activities in fragile ecosystem. [ABSTRACT FROM AUTHOR]

• Detailed methodology for reporting greenhouse gas emissions from organic soils. • Based on more than 250 annual GHG balances measured in Germany. • First approach to use representative water table distributions per land-use class. • Non-linear response functions for CO 2 and CH 4 are developed. • German drained organic soils emit more than 55 Mt CO 2-eq. yr−1, of which 90% are CO 2. Drained organic soils are large sources of anthropogenic greenhouse gases (GHG) in many European and Asian countries. Therefore, these soils urgently need to be considered and adequately accounted for when attempting to decrease emissions from the Agriculture and Land Use, Land Use Change and Forestry (LULUCF) sectors. Here, we describe the methodology, data and results of the German approach for measurement, reporting and verification (MRV) of anthropogenic GHG emissions from drained organic soils and outline ways forward towards tracking drainage and rewetting. The methodology was developed for and is currently applied in the German GHG inventory under the United Nations Framework Convention on Climate Change (UNFCCC) and the Kyoto Protocol. Spatial activity data comprise high resolution maps of land-use, type of organic soil and mean annual water table (WT). The WT map was derived by a boosted regression trees model from data of more than 1000 dipwells. Emissions of carbon dioxide (CO 2), nitrous oxide (N 2 O) and methane (CH 4) were synthesized from a unique national data set comprising more than 250 annual GHG balances from 118 sites in most land-use categories and types of organic soils. Measurements were performed with harmonized protocols using manual chambers. Non-linear response functions describe the dependency of CO 2 and CH 4 fluxes on mean annual WT, stratified by land-use where appropriate. Modelling results were aggregated into "implied emission factors" for each land-use category, taking into account the uncertainty of the response functions, the frequency distribution of the WT within each land-use category and further GHG sources such as dissolved organic carbon or CH 4 emissions from ditches. IPCC default emission factors were used for these minor GHG sources. In future, response functions could be applied directly when appropriate WT data is available. As no functional relationship was found for N 2 O emissions, emission factors were calculated as the mean observed flux per land-use category. In Germany, drained organic soils emit more than 55 million tons of GHGs per year, of which 91% are CO 2. This is equivalent to around 6.6% of the national GHG emissions in 2014. Thus, they are the largest GHG source from agriculture and LULUCF. The described methodology is applicable on the project scale as well as in other countries where similar data are collected. [ABSTRACT FROM AUTHOR]

• Framework for generating marine species models transferable in time. • Objective grouping of species to deal with a high number effectively. • Evaluation of nine algorithms and definition of the most transferable. • Machine learning algorithms are more sensitive than regression-based approaches. • Time and cost-effective tool for climate change studies in large marine areas. Species distribution models (SDMs 1 Species Distribution Models (SDM) 1) are crucial for guiding management in a changing world. However, limited understanding of algorithm selection, ensemble weights and temporal transferability assessment undermines confidence in their predictions. Transferable predictive models, based on objective and proven selection criteria could therefore provide effective tools for defining species-environment relationships. This study developed a framework for generating SDMs in the marine environment that improves models' temporal transferability. The methodological approach steps were: 1) Collection of predictors related to species ecology and their records and species grouping according to their ecological requirements. Twenty-one seaweeds were used as a case study. Environmental and distribution data were divided into two independent periods to evaluate temporal transferability. 2) A model for each species was built in each period with nine algorithms (Generalized Linear Model, Generalized Additive Model, Multivariate Adaptive Regression Spline, Mixture Discriminant Analysis, Classification and Regression Trees, Support Vector Machine, Flexible Discriminant Analysis, Random Forest, MAXENT) and projected into the other period. Predictor contributions to the final models were obtained. 3) Assessment of predictive performance for each model was made using the area under the receiver operating characteristic curve and true skill statistics metric for both models' accuracy and temporal transferability capabilities. All values were over 0.8 for all groups. In turn, the geographical pattern of all models were shown to be ecologically coherent. The algorithms and their weights that fit best were used to generate transferable models over time in the marine environment and retained for each species. In general, machine learning algorithms produce models with higher sensitivity than regression-based approaches. This methodology sets the scene for further inquiries in the marine environment when developing consistent practices for model development and transferability. Results are satisfactory for broad application in marine research, allowing a comparative framework between species predictions and facilitating the use of transferable models, especially in climate change studies across large areas. In addition, the proposed methodological approach is a cost-effective tool for dealing with a high number of species in marine environments. All data are freely available, so the methodology can be reproduced for marine researchers with different objectives. [ABSTRACT FROM AUTHOR]