DSOLMap, a novel high-resolution global digital soil property map for the SWAT + model: Development and hydrological evaluation

CATENA, Journal Year: 2023, Volume and Issue: 231, P. 107339 - 107339

Published: June 28, 2023

This research paper addresses the ongoing challenge of developing fine-resolution global digital soil property maps for hydrological modelling applications. Hydrological models are essential understanding watershed dynamics and impact human activities on water resources. Soil data, which plays a crucial role in cycle, is requisite model input. Global usually have coarse spatial resolutions, adding considerable uncertainty to despite calibration efforts. To address this issue, new map with 250 m resolution, known as Digital Open Land Map (DSOLMap), was developed evaluated study. The DSOLMap has finer resolution than existing more detailed profile divided into six horizons. high-resolution tailored SWAT + format. latest released version Water Assessment Tool (SWAT), one most comprehensive models, widely used worldwide. A evaluation conducted its results were compared two other databases using basin located north Spain. findings showed that detailed, finer-resolution such those offers, improved performance daily scale before after validation procedures. represents step forward modelling, notably regions scarce or unavailable information. can help decision-makers challenges related resources environmental issues through modelling.

Language: Английский

---

Predictive performance of machine learning model with varying sampling designs, sample sizes, and spatial extents

Ecological Informatics, Journal Year: 2023, Volume and Issue: 78, P. 102294 - 102294

Published: Sept. 11, 2023

Language: Английский

---

Historical and future variation of soil organic carbon in China

Geoderma, Journal Year: 2023, Volume and Issue: 436, P. 116557 - 116557

Published: June 12, 2023

Revealing historical changes in soil organic carbon (SOC) and exploring its future status are important for safeguarding health food security, giving full play to the service function of ecosystems, coping with climate change. However, there is still a gap our understanding SOC stocks China their spatial patterns response Therefore, we attempted fill this knowledge using large amount observation data, digital mapping technology, global circulation models from Coupled Model Inter-comparison Project phase 6 (CMIP6). In study, random forest model was selected construct relationship between top 0–20 cm (SOC020) 0–100 (SOC0100) 21 environmental factors. Spatiotemporal 1980 2100 were revealed at resolution 1 km five-year interval three scenarios CMIP6. The cross-validation results indicated acceptable predictions both depths SOC; however, relatively prediction uncertainties observed SOC0100 Tibetan Plateau northeastern China. mean values SOC020 over last four decades 35.77 84.62 Tg, respectively, showed sinks national scale, accumulation rates 0.05 Pg yr-1 0.036 y-1 two depths. Compared (1980–2020), will fluctuate significantly under different scenarios. Among them, slow increasing trend SSP1-1.9 low emission scenario, while presented decreasing SSP2-4.5 SSP5-8.5 medium–high particular, most larger likelihood being source. This study provides reference pools change evaluating effectiveness land management ecological protection.

Language: Английский

---

Remote Sensing Data for Digital Soil Mapping in French Research—A Review

Remote Sensing, Journal Year: 2023, Volume and Issue: 15(12), P. 3070 - 3070

Published: June 12, 2023

Soils are at the crossroads of many existential issues that humanity is currently facing. a finite resource under threat, mainly due to human pressure. There an urgent need map and monitor them field, regional, global scales in order improve their management prevent degradation. This remains challenge high often complex spatial variability inherent soils. Over last four decades, major research efforts field pedometrics have led development methods allowing capture nature As result, digital soil mapping (DSM) approaches been developed for quantifying soils space time. DSM monitoring become operational thanks harmonization databases, advances modeling machine learning, increasing availability spatiotemporal covariates, including exponential increase freely available remote sensing (RS) data. The latter boosted DSM, resolution assessing changes through We present review main contributions developments French (inter)national research, which has long history both RS DSM. Thanks SPOT satellite constellation started early 1980s, communities pioneered using sensing. describes data, tools, imagery support predictions wide range properties discusses pros cons. demonstrates data frequently used (i) by considering as substitute analytical measurements, or (ii) covariates related controlling factors formation evolution. It further highlights great potential provides overview challenges prospects future sensors. opens up broad use natural monitoring.

Language: Английский

---

Towards Optimal Variable Selection Methods for Soil Property Prediction Using a Regional Soil Vis-NIR Spectral Library

Remote Sensing, Journal Year: 2023, Volume and Issue: 15(2), P. 465 - 465

Published: Jan. 12, 2023

Soil visible and near-infrared (Vis-NIR, 350–2500 nm) spectroscopy has been proven as an alternative to conventional laboratory analysis due its advantages being rapid, cost-effective, non-destructive environmentally friendly. Different variable selection methods have used deal with the high redundancy, heavy computation, model complexity of using full spectra in spectral modelling. However, most previous studies a linear algorithm selection, application non-linear remains poorly explored. To address current knowledge gap, based on regional soil Vis-NIR library (1430 samples), we evaluated seven algorithms together three predictive predicting properties. Our results showed that Cubist overperformed partial least squares regression (PLSR) random forests (RF) properties (R2 > 0.75 for organic matter, total nitrogen pH) when spectra. Most can greatly reduce number bands therefore simplified models without losing accuracy. The also there was no silver bullet optimal among different algorithms: (1) competitive adaptive reweighted sampling (CARS) always performed best PLSR algorithm, followed by forward recursive feature (FRFS); (2) elimination (RFE) genetic (GA) generally had better accuracy than others algorithm; (3) FRFS performance RF algorithm. In addition, matched outcome this study provides valuable reference information spectroscopic techniques algorithms.

Language: Английский

---