Prediction of Soil Particle Size Fraction using Geographically Weighted Regression and Random Forest

The development of spatial modeling for soil properties has progressed in recent decades. This responds to the growing demand for land spatial data and exact soil property prediction for agronomical reasons, particularly in precision farming, in order to speed up precision agricultural activities. In this regards a comparison of the GWR and RF models was carried out in order to determine which model is the best at forecasting surface soil texture and how dependable each model is at doing so. The purpose of this research is to get the best model in predicting particle soil fraction (PSF). 50 topsoil samples were collected from several locations in the Kalikonto Watershed, Indonesia, and the soil PSF (sand, silt, and clay) in the upper 10 cm varied. The LMV, slope, and elevation were calculated using DEM data and utilized as predictor variables. As a result, the weighting of the GWR model has a considerable impact on the final model, and all other factors have a major effect on the PSF determination. The RF, on the other hand, looks to be superior than the GWR variants. The RF model outperformed the other models in every PSF variable. This study reveals that topsoil quality and terrain attributes are linked, which may be assessed using field measurements and model projections. More research is needed to generate more efficient input parameters that will help with soil variability precision and accuracy of soil map products.

Use of Multi-Seasonal Satellite Images to Predict SOC from Cultivated Lands in a Montane Ecosystem

Although algorithms are well developed to predict soil organic carbon (SOC), selecting appropriate covariates to improve prediction accuracy is an ongoing challenge. Terrain attributes and remote sensing data are the most common covariates for SOC prediction. This study tested the predictive performance of nine different combinations of topographic variables and multi-season remotely sensed data to improve the prediction of SOC in the cultivated lands of a middle mountain catchment of Nepal. The random forest method was used to predict SOC contents, and the quantile regression forest for quantifying the prediction uncertainty. Prediction of SOC contents was improved when remote sensing data of multiple seasons were used together with the terrain variables. Remote sensing data of multiple seasons capture the dynamic conditions of surface soils more effectively than using an image of a single season. It is concluded that the use of remote sensing images of multiple seasons instead of a snapshot of a single period may be more effective for improving the prediction of SOC in a digital soil mapping framework. However, an image with the right timing of cropping season can provide comparable results if a parsimonious model is preferred.

Evaluating tree growth factors into species-specific functional soil maps for improved agroforestry system efficiency

Agroforestry systems play an important role in sustainable agroecosystems. However, accurately and adequately quantifying the relationships between environmental factors and tree growth in these systems are still lacking. Objectives of this study were to quantify environmental factors affecting growth of four tree species and to develop functional soil maps (FSM) for each species in an agroforestry site. The diameter at breast height, absolute growth rate (AGR), and neighborhood competition index of 259 trees from four species (northern red oak [Quercus rubra], pecan [Carya illinoinensis], cottonwood [Populus deltoides], and sycamore [Platanus occidentalis]) were determined. A total of 51 topsoil samples were collected and analyzed, and 12 terrain attributes were derived from the digital elevation model. The relationships between AGR, soil, topography, and tree size were analyzed using Spearman correlation. Based on correlation analysis, FSM for each species were generated using the k-means cluster method by overlaying correlated soil and terrain attribute maps. Results showed tree size and terrain attributes were driving factors affecting tree growth rate relative to soil properties. The spatial variations in AGR among functional units were statistically compared within tree species and the areas with larger AGR were identified by the FSM. This study demonstrated that FSM could delineate areas with different AGR for the oak, cottonwood, and sycamore trees. The AGR of pecan trees did not vary among functional units. The generated FSM may allow land managers to more precisely establish and manage agroforestry systems.

Modeling And Mapping of Soil Salinity And Alkalinity Using Remote Sensing Data And Topographic Factors

Soil salinity and alkalinity seriously threaten crop production, soil productivity and sustainable agriculture, especially in arid and semi-arid areas, leading to land degradation, therefore, spatial distribution of these parameters are really important for successful management of such areas. The surface soil salinity and sodium adsorption ratio (SAR) have been modeled in this article. Auxiliary data were terrain attributes derived from digital elevation model (DEM), remote sensing spectral bands, and indices of vegetation and salinity derived from Landsat 8 OLI satellite. In total, 118 soil samples were collected from depth of 0-15 cm in homogenous units at Doviraj plain in the southern part of Ilam province, western Iran. Saturated electrical conductivity (ECe), SAR and other soil properties were analyzed and calculated. To model ECe and SAR parameters with the auxiliary data, stepwise multi linear regression (SMLR) and random forest (RF) regression were applied. The highest accuracy were obtained through RF model with validation coefficient of determination (R2val) =0.82 and 0.83 and validation root mean square error (RMSEval)=7.40 dS/m and 11.20 for ECe and SAR respectively. Furthermore, results indicated that strongest influence on the prediction of soil salinity followed by Band10, principal component analysis (PC3), Vertical Distance to Channel Network (VDCN) and Analytical Hill Shading (AH). Also, Band10, Band11, Flow Accumulation (FA) and Topographic Wetness Index (TWI) were the important covariate in alkalinity prediction through RF model. Finally, it is suggested that similar techniques can be used to map and monitor soil salinity and alkalinity in other parts of arid regions.

Regionalization of hydrological model parameters using gradient boosting machine

Regionalization of hydrological model parameters is key to hydrological predictions in ungauged basins. The commonly used multiple linear regression (MLR) method may not be applicable in complex and nonlinear relationships between model parameters and watershed properties. Moreover, most regionalization methods assume lumped parameters for each catchment without considering within-catchment heterogeneity. Here we incorporated the Penman-Monteith-Leuning (PML) equation into the Distributed Time-Variant Gain Model (DTVGM) to improve the mechanistic representation of the evapotranspiration process. We calibrated six key model parameters grid-by-grid across China using a multivariable calibration strategy, which incorporates spatiotemporal runoff and evapotranspiration (ET) datasets (0.25°, monthly) as reference. In addition, we used the gradient boosting machine (GBM), a machine learning technique, to portray the dependence of model parameters on soil and terrain attributes in four distinct climatic zones across China. We show that the modified DTVGM could reasonably estimate the runoff and ET over China using the calibrated parameters, but performed better in humid than arid regions for the validation period. The regionalized parameters by the GBM method exhibited better spatial coherence relative to the calibrated grid-by-grid parameters. In addition, GBM outperformed the stepwise MLR method in both parameter regionalization and gridded runoff simulations at national scale, though the improvement is not significant pertaining to watershed streamflow validation due to most of the watersheds being located in humid regions. We also revealed that the slope, saturated soil moisture content, and elevation are the most important explanatory variables to inform model parameters based on the GBM approach. The machine-learning-based regionalization approach provides an effective alternative to deriving hydrological model parameters by using watershed properties in ungauged regions.

The influence of subcolony-scale nesting habitat on the reproductive success of Adélie penguins

Group-size variation is common in colonially breeding species, including seabirds, whose breeding colonies can vary in size by several orders of magnitude. Seabirds are some of the most threatened marine taxa and understanding the drivers of colony size variation is more important than ever. Reproductive success is an important demographic parameter that can impact colony size, and it varies in association with a number of factors, including nesting habitat quality. Within colonies, seabirds often aggregate into distinct groups or subcolonies that may vary in quality. We used data from two colonies of Adélie penguins 73 km apart on Ross Island, Antarctica, one large and one small to investigate (1) How subcolony habitat characteristics influence reproductive success and (2) How these relationships differ at a small (Cape Royds) and large (Cape Crozier) colony with different terrain characteristics. Subcolonies were characterized using terrain attributes (elevation, slope aspect, slope steepness, wind shelter, flow accumulation), as well group characteristics (area/size, perimeter-to-area ratio, and proximity to nest predators). Reproductive success was higher and less variable at the larger colony while subcolony characteristics explained more of the variance in reproductive success at the small colony. The most important variable influencing subcolony quality at both colonies was perimeter-to-area ratio, likely reflecting the importance of nest predation by south polar skuas along subcolony edges. The small colony contained a higher proportion of edge nests thus higher potential impact from skua nest predation. Stochastic environmental events may facilitate smaller colonies becoming “trapped” by nest predation: a rapid decline in the number of breeding individuals may increase the proportion of edge nests, leading to higher relative nest predation and hindering population recovery. Several terrain covariates were retained in the final models but which variables, the shapes of the relationships, and importance varied between colonies.