Geographically Weighted Regression Kriging Research Articles

Influences of Snow Cover on the Thermal Regimes of Xing'an Permafrost in Northeast China in 1960s–2010s

ABSTRACTThe distributive characteristics of snow cover and their impacting mechanisms on ground thermal regimes in Northeast China remain evasive because of limited systematic studies. In this study, based on long‐term ground‐based observational data and auxiliary topographic data, geographically weighted regression kriging (GWRK) method and the temperature at the top of permafrost (TTOP) model were used to analyze the influences of snow cover duration (SCD) and average snow depth over the SCD (ASDSCD) on the thermal regimes of Xing'an permafrost in Northeast China in the 1960s–2010s. The results show a remarkable reduction of permafrost extent in Northeast China from the 1960s to 2010s, with an average reduction rate of 4.115 × 104 km2/decade. In permafrost regions, from the 1960s to 2010s, average SCD varied between 150 and 160 days, and the regional average of ASDSCD between 8 and 14 cm. Increases in ASDSCD led to a rise of TTOP. From the 1960s to 2010s, the regional average of ASDSCD increased by 3.53 cm, and that of TTOP by 2.02°C. The research results can provide scientific basis and data support for evaluating the responses of permafrost and cold region ecosystems to climate change in Northeast China.

A multiscale geographically weighted regression kriging method for spatial downscaling of satellite-based ozone datasets

Accurate monitoring of ozone (O3) concentrations by remote sensing is essential for achieving pollution control and ecological protection. However, the existing O3 remote sensing data with a low spatial resolution do not facilitate fine-grained studies of small-scale urban clusters. In this study, the multiscale geographically weighted regression kriging (MGWRK) method was used to spatially downscale O3 remote sensing products (10 km × 10 km). Downscaling factors were selected from meteorological factors and vegetation, aerosol optical thickness (AOD), and air pollutant emission inventory data. Spatial heterogeneity and scale differences among the factors were considered and compared via multiple regression kriging (MLRK) and geographically weighted regression kriging (GWRK) to generate 1-km annual and seasonal O3 remote sensing products. The results showed that I) the downscaling accuracy of each model can be expressed as MGWRK > GWRK > MLRK; the local downscaling model yields data that are more consistent with the actual spatial distribution of O3 after considering the spatial heterogeneity of the influencing factors; and the downscaled annual and seasonal data exhibit satisfactory spatial texture characteristics and consistency with the original spatial distribution of O3, while the distribution boundary problem of image elements is eliminated. II) Nitrogen oxide (NOx) and volatile organic compound emissions and temperature exhibit strong positive correlations with O3, while wind speed, humidity, the normalized difference vegetation index, and AOD indicate weak positive correlations with O3. Moreover, precipitation exhibits a weak negative correlation with O3. III) The coefficient of determination (R2) of the 1-km resolution annual O3 concentration data after downscaling based on the MGWRK model reaches 0.93, while the RRMSE and MAE values are only 3% and 1.86, respectively, with a coefficient of variation of 9.55%; the downscaling accuracy of the seasonal O3 concentration data is higher in summer and winter than during the other seasons, with R2 greater than 0.85, further confirming the spatial and temporal downscaling advantages of the MGWRK model for O3 in the Chang-Zhu-Tan city cluster. This further corroborates the feasibility of the MGWRK model for spatial and temporal O3 downscaling in the Chang-Zhu-Tan urban area.

Regional-Scale Topsoil Organic Matter Estimation Based on a Geographic Detector Model Using Landsat Data, Pingtan Island, Fujian, China

Understanding the spatial distribution of soil organic matter (SOM) is important for land use management, but conventional sampling methods require significant human and financial resources. How to map SOM and monitor its changes using a limited number of sample points combined with remote sensing techniques that provide long-time series data is crucial. This study aimed to generate a regional-scale near-surface SOM map using 70 soil samples and covariate environmental factors extracted mainly from Landsat 8 OLI. Firstly, the sensitivity of each environmental factor to SOM was tested using a geographic detector model (GDM). Secondly, the tested factors were selected for modeling and mapping by ordinary least squares (OLS) and geographically weighted regression kriging (GWRK). The performance of these two models was compared. Finally, the mapping results of the better model (GWRK) were compared and analyzed with the traditional interpolation results based solely on sampling points to verify the rationality of the proposed method. The results show that three environmental factors, ratio vegetation index (RVI), differential vegetation index (DVI), and terrain roughness (TR), have a strong influence on the spatial variability of SOM. Using these three factors in combination with the GWRK method, a more accurate and refined spatial distribution map of SOM can be obtained. Comparing the SOM maps of GWRK and the traditional interpolation method, the results show that the accuracy of GWRK (R2 = 0.405; mean absolute error = 0.637, and root mean square error = 0.813) is higher than that of traditional interpolation methods (R2 = 0.291, MAE = 0.609, and RMSE = 0.863). The spatial recognition rate (fineness) of SOM patches at all levels using the GWRK method increased by more than 73 times compared to the traditional kriging. We conclude that the combination of limited SOM samples, environmental variables, GDM, and GWRK is a pragmatic approach for estimating regional-scale SOM.

Mapping Waterlogging Damage to Winter Wheat Yield Using Downscaling–Merging Satellite Daily Precipitation in the Middle and Lower Reaches of the Yangtze River

Excessive water and water deficit are two important factors that limit agricultural development worldwide. However, the impact of waterlogging on winter wheat yield on a large scale, compared with drought caused by water deficit, remains unclear. In this study, we assessed the waterlogging damage to winter wheat yield using the downscaled and fused TRMM 3B42 from 1998 to 2014. First, we downscaled the TRMM 3B42 with area-to-point kriging (APK) and fused it with rain gauge measurements using geographically weighted regression kriging (GWRK). Then, we calculated the accumulated number of rainy days (ARD) of different continuous rain processes (CRPs) with durations ranging from 5 to 15 days as a waterlogging indicator. A quadratic polynomial model was used to fit the yield change rate (YCR) and the waterlogging indicator, and the waterlogging levels (mild, moderate, and severe) based on the estimated YCR from the optimal model were determined. Our results showed that downscaling the TRMM 3B42 using APK improved the limited accuracy, while GWRK fusion significantly increased the precision of quantitative indicators, such as R (from 0.67 to 0.84), and the detectability of precipitation events, such as the probability of detection (POD) (from 0.60 to 0.78). Furthermore, we found that 67% of the variation in the YCR could be explained by the ARD of a CRP of 11 days, followed by the ARD of a CRP of 13 days (R2 of 0.65). During the typical wet growing season of 2001–2002, the percentages of mild, moderate, and severe waterlogged pixels were 5.72%, 2.00%, and 0.63%, respectively. Long time series waterlogging spatial mapping can clearly show the distribution and degree of waterlogging, providing a basis for policymakers to carry out waterlogging disaster prevention and mitigation strategies.

CLASSIFICATION OF STUNTING USING GEOGRAPHICALLY WEIGHTED REGRESSION-KRIGING CASE STUDY: STUNTING IN EAST JAVA

Geographically Weighted Regression Kriging (GWRK) is a special case of Geographically Weighted Regression (GWR) model, which is modeling with the effect of spatial autocorrelation on the GWR model error. The purpose of this research is to obtain a GWRK model between the factors that affect stunting density for each site viewed from the district center point in East Java Province and to make a prediction map based on the GWRK modeling. The data used was obtained from Basic Health Research (RISKESDAS) and the East Java Health Profile Book for 2021. The units of observation in this study were 38 districts in East Java.. Based on the GWR modeling results, it was found that the GWR model error contained spatial autocorrelation so that GWR model can be formed. From the GWRK modeling using stunting prevalence data in East Java in 2021, it was found that the GWR model was better than the global regression. Through prediction and prediction mapping formed from the GWR-Kriging modeling, it could be seen that stunting in regencies in East Java was evenly distributed . The interpolation map showed that the stunting forecasting values using the Kriging GWR interpolation ranged from 27% to 46%.

Spatial distribution and influencing factors of humus layer thickness of forest land in permafrost region of Northeast China

Humus has significant impacts on soil water dynamics, nutrient storage and plant growth. Studies have demonstrated that the humus layer is related to climate, soil properties, topography and other environmental factors. However, the thickness of the humus layer (HLT) and its spatial distributions are less investigated. Especially in boreal mountain forests underlain with permafrost, the environmental influences on HLT remain unclear. In a high-latitude watershed of boreal mountain forest in Northeast China, this study explores the effects of the environmental factors on HLT, and its relationships with permafrost types. The HLT distribution across the watershed is extracted using the geographically weighted regression kriging (GWRK) model. The elevation, slope, aspect, mean annual ground surface temperature and mean annual soil water content are determined to be statistically significant influencers factors of HLT. The elevation is the most important variable for HLT in this cold land, with an indirect effect on HLT through impacting ground surface temperature and soil water content. Topography casts more significant impact than ground surface temperature and soil water content. HLT ranges from 1.1 to 26.7 cm across three permafrost classes in the study area, decreasing with the discontinuity of permafrost types (continuous permafrost > sporadic permafrost > isolated patches). The findings suggest that the humus development is closely related to the permafrost type in northern high latitudes, and the permafrost degradation can lead to humus layer thinning under climate warming.

Spatial Variation and Influencing Factors of Trace Elements in Farmland in a Lateritic Red Soil Region of China

Trace elements in farmland soil are important indicators of soil quality and farmland health, and also maintain the nutrient balance and promote the healthy growth of plants. In this study, taking Conghua District of Guangzhou city as the study area, the effects of topography, soil, land use, and other factors on trace elements in soil were investigated, and the spatial variability of boron (B), manganese (Mn), molybdenum (Mo), copper (Cu), and zinc (Zn) in farmland soil in a typical red soil region were mapped using a geographically weighted regression (GWR) method. The pH and land economic index (LEI) were important factors affecting the changes in trace element concentrations in the five soils, and the Cu and Zn concentrations were clearly affected by human factors. In the study area, 86.99% of B measurements were classified as low and very low levels, 50.61% and 49.20% of Mo measurements were also low and very low, 71.79% of Mn measurements were classified as moderate, while 91.02% of Cu and 52.95% of Zn measurements were classified as high. After a cross validation, the GWR Kriging (GWRK) model results of each element were relatively stable, and the order of the fitting coefficient (R2) was Cu > Zn > B> Mn > Mo. This study clarifies the spatial distribution and influencing factors of soil microelements in the studied region. This information can be used to improve the nutrient imbalance, further guide agricultural production, strengthen the management of farmland, and improve the healthy productivity of cultivated land.

Contrasting multiple deterministic interpolation responses to different spatial scale in prediction soil organic carbon: A case study in Mollisols regions

Accurate estimation of SOM content of Mollisols influences fundamental sub-micron to global-scale biogeochemical processes and carbon-climate feedbacks. Comparative analysis of the responses of Multiple deterministic Interpolation models to spatial scale variations is the basis for multiscale soil organic carbon (SOC) pools simulation and evaluation. This study mainly manifested as spatial variation at different analytical levels of spatial correlation and the change of characteristic data attributes that characterize this change. Three spatial scales were selected (HQ, endemic area; XF, local area; JX, most area). A set of 164 topsoils (0–20 cm, sampling density of 0.1 points / km2) samples were taken, and 14 environmental variables and 14 soil characters variables were employed to contrast prediction accuracy of ordinary kriging (OK), inverse distance weighting (IDW), radial basis function (RBF), global polynomial (GPI), local polynomial (LPI), regression kriging interpolation methods such as (RK) and geographically weighted regression kriging (GWRK) responses to different spatial scale. Using a cross-validation procedure to evaluate the performance of the models. Overall, RK and GWRK, due to the introduction of the auxiliary variables, effectively predict SOM content and the improvement of prediction accuracy depends on the spatial scale, compared with IDW, RBF, GPI, and LPI.Furthermore, soil characters variables (TN, pH) as covariates have a more significant impact on the final prediction than environmental variables (Elevation, Slope, etc.). Therefore, the mutations and trends associated with the spatial resolution are analyzed in different spaces. The root means square error (RMSE) of RK and GWRK are reduced by 46.66 % and 64.26 % compared with OK. RK with fixed the regression coefficient obtained by OLS reveals that the uniformity of the object distribution in HQ and XF have Adjusted R2 0.955 and 0.915. Otherwise, the RI of GWRK with 0.792 adjusted R2 in JX is relatively 17.6 % higher than RK. The results from our case study highlight a key role in selecting the optimal spatial interpolation method for a given dataset associated with spatial scales and analyzing their regularity and applicability on different spatial scales.

Integration of a process-based model into the digital soil mapping improves the space-time soil organic carbon modelling in intensively human-impacted area

Soil organic carbon (SOC) in intensively human-impacted areas often presents extremely complicated patterns in space and time, which brought a new challenge to space–time SOC modelling. Here we attempt to create a space-for-time GWRK-RothC model by incorporating the predictions of the process-based RothC model into the geographically weighted regression kriging (GWRK) for improving the space–time modelling of SOC in intensively human-impacted area. We collected a total of 1219 topsoil samples (0–20 cm) from southern Jiangsu Province of China in 1980, 2000, and 2015, and modelled the spatiotemporal dynamics of SOC by using the modified RothC model (with a newly-added pH modifier for representing the impact of soil acidification on SOC decomposition), the space-for-time GWRK, and the proposed space-for-time GWRK-RothC model. Results showed that the change trends in SOC derived by the three methods were similar, with an overall trend of SOC increase during 1980–2000 and SOC decline in the subsequent 15 years (2000–2015). However, their performances for space–time SOC modelling were different. The proposed space-for-time GWRK-RothC model had the highest accuracy in representing the spatiotemporal evolution of SOC in the study area, with the lowest root mean square error (RMSE) (3.06 g kg−1), mean error (ME) (-0.19 g kg−1), and the highest coefficient of determination (R2) (0.61), compared to the accuracy of space-for-time GWRK (RMSE = 3.53 g kg−1, ME = -0.36 g kg−1, and R2 = 0.45) and RothC model (RMSE = 4.23 g kg−1, ME = 0.20 g kg−1, and R2 = 0.20). These results highlighted the importance of integrating process-model-derived SOC spatiotemporal dynamics information in space-for-time digital soil mapping (DSM) for improving the space–time SOC modelling. We believe that outputs from this study may offer valuable guidance for developing space–time DSM model to achieve better performance of SOC modelling in space and time, especially for areas affected by strong anthropogenic activities.

Prediction of Soil Heavy Metal Distribution Using Geographically Weighted Regression Kriging.

Both soil heavy metals and the influencing factors are related to spatial location and are spatially heterogeneous. However, the global linear regression model assumes the regression coefficients to be spatially stationary throughout the study region and is unable to account for the spatially varying relationships between soil heavy metals and influencing factors. Thus, the objectives of this study were to estimate the spatial distribution of soil heavy metals using a geographically weighted regression kriging (GWRK) approach, and compare the GWRK results with those obtained from ordinary kriging (OK) and regression kriging (RK). A dataset of soil lead (Pb) concentrations in Daye city, China, that was sampled in 2019 was used. According to the results of spatial smoothness, variability, and interpolation accuracy, GWRK was the best method and could provide the most reasonable spatial distribution pattern and the highest spatial interpolation accuracy in comparison with OK and RK.

High spatiotemporal resolution mapping of PM2.5 concentrations under a pollution scene assumption

Accurate individual exposure assessment to particulates in complex urban environments requires maps of PM2.5 concentration at high spatiotemporal resolution. Previous empirical researches of PM2.5 mapping usually have ignored the contextual influences of associated factors on pollution variation. This study presents a new thinking about spatial prediction of PM2.5 pollution based on the pollution scene assumption. Methodologically, pollution scenes are areas exert contextual influences on the spatiotemporal variety of air pollution and can be expressed by urban microenvironment dependence and temporal nonstationarity. Taking Changsha, China as a case, a two-stage modelling strategy of geographically weighted regression kriging (GWRK) was developed to validate the assumption based on a high-density sampling campaign and a fine-scale, manually interpreted urban microenvironment map. Our results confirm the potential existence of urban air pollution scene. PM2.5 concentration varies between urban microenvironments; pollution scene based GWRK is effective for high resolution mapping of PM2.5 concentration at the hourly scale and depicts more detailed spatial variations than traditional GWR in this study. This assumption and modelling strategy provide a promising way for mapping urban air pollution at high resolution which will further benefits works on exposure assessment and risk avoidance at fine scales.

Ingestion of GNSS-Derived ZTD and PWV for Spatial Interpolation of PM2.5 Concentration in Central and Southern China.

With the increasing application of global navigation satellite system (GNSS) technology in the field of meteorology, satellite-derived zenith tropospheric delay (ZTD) and precipitable water vapor (PWV) data have been used to explore the spatial coverage pattern of PM2.5 concentrations. In this study, the PM2.5 concentration data obtained from 340 PM2.5 ground stations in south-central China were used to analyze the variation patterns of PM2.5 in south-central China at different time periods, and six PM2.5 interpolation models were developed in the region. The spatial and temporal PM2.5 variation patterns in central and southern China were analyzed from the perspectives of time series variations and spatial distribution characteristics, and six types of interpolation models were established in central and southern China. (1) Through correlation analysis, and exploratory regression and geographical detector methods, the correlation analysis of PM2.5-related variables showed that the GNSS-derived PWV and ZTD were negatively correlated with PM2.5, and that their significances and contributions to the spatial analysis were good. (2) Three types of suitable variable combinations were selected for modeling through a collinearity diagnosis, and six types of models (geographically weighted regression (GWR), geographically weighted regression kriging (GWRK), geographically weighted regression—empirical bayesian kriging (GWR-EBK), multiscale geographically weighted regression (MGWR), multiscale geographically weighted regression kriging (MGWRK), and multiscale geographically weighted regression—empirical bayesian kriging (MGWR-EBK)) were constructed. The overall R2 of the GWR-EBK model construction was the best (annual: 0.962, winter: 0.966, spring: 0.926, summer: 0.873, and autumn: 0.908), and the interpolation accuracy of the GWR-EBK model constructed by inputting ZTD was the best overall, with an average RMSE of 3.22 μg/m3 recorded, while the GWR-EBK model constructed by inputting PWV had the highest interpolation accuracy in winter, with an RMSE of 4.5 μg/m3 recorded; these values were 2.17% and 4.26% higher than the RMSE values of the other two types of models (ZTD and temperature) in winter, respectively. (3) The introduction of the empirical Bayesian kriging method to interpolate the residuals of the models (GWR and MGWR) and to then correct the original interpolation results of the models was the most effective, and the accuracy improvement percentage was better than that of the ordinary kriging method. The average improvement ratios of the GWRK and GWR-EBK models compared with that of the GWR model were 5.04% and 14.74%, respectively, and the average improvement ratios of the MGWRK and MGWR-EBK models compared with that of the MGWR model were 2.79% and 12.66%, respectively. (4) Elevation intervals and provinces were classified, and the influence of the elevation and the spatial distribution of the plane on the accuracy of the PM2.5 regional model was discussed. The experiments showed that the accuracy of the constructed regional model decreased as the elevation increased. The accuracies of the models in representing Henan, Hubei and Hunan provinces were lower than those of the models in representing Guangdong and Guangxi provinces.

Spatial distribution of carbon storage in natural secondary forest based on geographically weighted regression expansion model.

To accurately assess carbon storage and its spatial distribution in natural secondary forest at the regional scale, we constructed seven expansion models by modifying the geographically weighted regression (GWR) in aspects of spatial dimension, parameter heterogeneity and residual spatial autocorrelation, based on data collected from 165 bureau level permanent plots in Langxi Forest Farm of Wangqing Forestry Bureau in Jilin Province. Stand factor, topography factor, and soil factor were selected as the influencing factors. The expansion models included geographically and altitudinal weighted regression (GAWR), semiparametric geographically weighted regression (SGWR), semiparametric geographically and altitudinal weighted regression (SGAWR), geographically weighted regression Kriging (GWRK), geographically and altitudinal weighted regression Kriging (GAWRK), semiparametric geographically weighted regression Kriging (SGWRK), and semiparametric geographically and altitudinal weighted regression Kriging (SGAWRK). Coefficient of determination (R2), mean square error (MSE) and Akaike's Information Criterion (AIC) were used to evaluate the fitness of these models. Finally, the spatial distribution diagram of forest carbon storage was drawn with the fitting results of the optimal regression model, and the distribution pattern of forest carbon storage in the research area was analyzed. The stand factor and topographic factor had strong influence on carbon storage of natural secondary forests, among which the average diameter at breast height (DBH) of stands was the dominant variable. There was positive correlation between stand factor and topographic factor. SGWR and SGAWR model could reduce the spatial autocorrelation of the GWR model residual. The geographically regression expansion model could improve the fitting effect of GWR model. Among them, the SGWRK model had the highest R2 and the lowest MSE and AIC. The method with altitude as the spatial weight did not effectively improve the fitting effect of the model. The total forest carbon storage of Langxi Forest Farm was 205×104 t, and the carbon density ranged from 8.56 to 145.74 t·hm-2, with a mean value of 57.98 t·hm-2. Overall, the distribution pattern of carbon storage was high in the northwest and low in the southeast, while high in the edge and low in the interior. By improving the parameter heterogeneity and residual spatial autocorrelation in the GWR model, we can accurately assess the spatial relationship between forest carbon storage and relevant variables in the study area, and improve the estimation accuracy of the forest carbon storage and its spatial distribution at the regional scale.

Comparative Study on Spatial Digital Mapping Methods of Soil Nutrients Based on Different Geospatial Technologies

Soil organic matter (SOM), total nitrogen (TN), available phosphorus (AP), and available potassium (AK) are important indicators of soil fertility when undertaking a quality evaluation. Obtaining a high-precision spatial distribution map of soil nutrients is of great significance for the differentiated management of nutrient resources and reducing non-point source pollution. However, the spatial heterogeneity of soil nutrients lead to uncertainty in the modeling process. To determine the best interpolation method, terrain, climate, and vegetation factors were used as auxiliary variables to participate in the investigation of soil nutrient spatial modeling in the present study. We used the mean error (ME), mean absolute error (MAE), root mean square error (RMSE), and accuracy (Acc) of a dataset to comprehensively compare the performance of four different geospatial techniques: ordinary kriging (OK), regression kriging (RK), geographically weighted regression kriging (GWRK), and multiscale geographically weighted regression kriging (MGWRK). The results showed that the hybrid methods (RK, GWRK, and MGWRK) could improve the prediction accuracy to a certain extent when the residuals were spatially correlated; however, this improvement was not significant. The new MGWRK model has certain advantages in reducing the overall residual level, but it failed to achieve the desired accuracy. Considering the cost of modeling, the OK method still provides an interpolation method with a relatively simple analysis process and relatively reliable results. Therefore, it may be more beneficial to design soil sampling rationally and obtain higher-quality auxiliary variable data than to seek complex statistical methods to improve spatial prediction accuracy. This research provides a reference for the spatial mapping of soil nutrients at the farmland scale.

Modeling and Predictive Mapping of Soil Organic Carbon Density in a Small-Scale Area Using Geographically Weighted Regression Kriging Approach

Different natural environmental variables affect the spatial distribution of soil organic carbon (SOC), which has strong spatial heterogeneity and non-stationarity. Additionally, the soil organic carbon density (SOCD) has strong spatial varying relationships with the environmental factors, and the residuals should keep independent. This is one hard and challenging study in digital soil mapping. This study was designed to explore the different impacts of natural environmental factors and construct spatial prediction models of SOC in the junction region (with an area of 2130.37 km2) between Enshi City and Yidu City, Hubei Province, China. Multiple spatial interpolation models, such as stepwise linear regression (STR), geographically weighted regression (GWR), regression kriging (RK), and geographically weighted regression kriging (GWRK), were built using different natural environmental variables (e.g., terrain, environmental, and human factors) as auxiliary variables. The goodness of fit (R2), root mean square error, and improving accuracy were used to evaluate the predicted results of the spatial interpolation models. Results from Pearson correlation coefficient analysis and STR showed that SOCD was strongly correlated with elevation, topographic position index (TPI), topographic wetness index (TWI), slope, and normalized difference vegetation index (NDVI). GWRK had the highest simulation accuracy, followed by RK, whereas STR was the weakest. A theoretical scientific basis is, therefore, provided for exploring the relationship between SOCD and the corresponding environmental variables as well as for modeling and estimating the regional soil carbon pool.

Effects of freeze-thaw cycles on the spatial distribution of soil total nitrogen using a geographically weighted regression kriging method

Freeze-thaw cycles (FTCs) change the soil physicochemical properties and biogeochemical cycles and possibly also change the spatial heterogeneity of soil total nitrogen (TN) in the watershed. In this study, 912 soil samples were collected at 0–5 cm, 5–10 cm and 10–20 cm soil depths in the autumn and the spring of next year after FTCs of 2016–2017 and 2017–2018 in a Mollisol watershed (1.86 km2) of northeast China. The field investigations combined with classical statistics and geographically weighted regression kriging (GWRK) were used to explore the spatiotemporal distribution of TN before and after FTCs. Terrain information (e.g., slope aspect) and land management (e.g., tillage method) was main covariates were used for GWRK. The results showed the following. (1) TN decreased by 3.7–5.7% after FTCs at 0–20 cm soil depths at the watershed scale, decreasing more than 60% of the total watershed area. (2) The spatial pattern of TN did not change in the field with slope aspects and tillage methods after FTCs, but it changed with slope steepness and land uses. (3) TN was mainly influenced by snowmelt erosion during FTCs. TN increased in parts of the top slope, at land use intersection, in gully banks and at the watershed outlet. (4) Simulation accuracy of GWRK was higher than ordinary kriging (OK) for predicted TN at 0–20 cm soil depths before and after FTCs. (5) Spatial distribution of soil TN after FTCs can be predicted (R2 = 0.521, p < 0.0001) and validated (R2 = 0.494, p < 0.0001) using the data before FTCs based on GWRK. Generally, to reduce N loss and increase farmland fertility after FTCs, conservational techniques, e.g., tillage and straw amendment, could be used, especially in the middle slope positions. Moreover, fertilization should be appropriately reduced in parts of the watershed after FTCs, especially on the top slope, land use intersection and watershed outlet.

Determination of Stunting Risk Factors Using Spatial Interpolation Geographically Weighted Regression Kriging in Malang

Stunting is the condition toddlers have Stunting is the condition toddlers have less length or height if compared to age. The high percentage of stunting is influenced by several factors, namely access to healthy latrines, quality drinking water, hand washing behavior with soap, coverage of posyandu access and coverage of breast milk 1-6 months, and there are indications that if an area has a high stunting percentage, then there is a possibility that the nearest area has the same condition. So, the statistic method for this research use the spatial interpolation Geographically Weighted Regression Kriging. Geographically Weighted Regression (GWR) is a weighted regression in which the weighting function is used to describe the closeness of relations between regions. The weight used is distance based weight dan weighting by area (contiguity). Ordinary kriging method calculated with semivariogram which is one function to describe and model the spatial autocorrelation between data of a variable and function as a measure of variance. The results showed that based on value GWR model with weight Fixed Gaussian Kernel better to use then the weighted GWR model Rook Contiguity. The Predicted of prevelensi stunting in the form of map based on interpolation GWR Kriging. Keywords: Stunting, GWR, and Kriging.

Spatial prediction and mapping of soil pH across a tropical afro-montane landscape

Soil pH is an indispensable part of the soil bionetwork, but so vulnerable to the dynamics of land-use change. Understanding the spatial and temporal distribution of soil pH is thus important for sustainable land use planning and management. In this study, we developed an approach based on geographically weighted regression-kriging (GWRK) to predict the continuous variation of soil pH up to 15 cm depth. The approach is compared to multiple linear regression-kriging (MLRK) using a total of 220 soil observations and 28 candidate auxiliary data, including climatic, topographic, and remotely sensed data. Results revealed that organic matter, sand, silt, temperature, and Landsat 8 operational land imager band 7 can explain the spatial variability of soil pH. GWR-based (local) models remarkably improved the prediction accuracy of soil pH compared to the MLR-based (global) models considering the root mean square error (i.e., 0.33 vs. 0.15; 0.37 vs. 0.17, respectively) and the coefficient of multiple determination (R2) (i.e., 0.59 vs. 0.74; 0.43 vs. 0.55, respectively). In conclusion, the use of GWR-based models revealed that the correlations between soil pH and environmental covariates were not stationary in space. Forested areas tended to be very strongly to moderately acidic, while croplands were slightly acidic to neutral. The GWR-based models demonstrated their utility in improving predictive soil mapping. These findings will support spatially-targeted soil pH management for improved food security and ecosystem health.

Mapping soil organic carbon and total nitrogen in croplands of the Corn Belt of Northeast China based on geographically weighted regression kriging model

Understanding the spatial distributions of soil organic carbon (SOC) and total nitrogen (TN) in croplands is necessary for the sustainable management of soil resources. In this study, a geographically weighted regression kriging (GWRK) model was applied to investigate spatial variabilities of SOC and TN in croplands of the Corn Belt of Northeast China (CBNC). Based on this, a map of the ratio of SOC and TN (C:N) was obtained. Results of geostatistical analysis showed that the best-fit variogram models for SOC and TN were spherical and exponential, respectively; the spatial dependence structures of SOC and TN were strong. The variabilities of SOC and TN were explained by GWRK model having a relatively high R2, and low root mean square error (RMSE) and mean absolute error (MAE) indices. These indices showed the potential applications of GWRK model on a large scale when environmental variables are quite different and the global trend is weak. The area under the curve (AUC) –receiver operating characteristic (ROC) curve exhibited that the prediction of SOC has a general performance whereas that of TN exhibited a good performance. Annual mean temperature had the largest impact on the prediction of SOC and TN in the CBNC. Mediated by topographic and climatic variables, remote sensing-derived variables did not play prominent roles as other studies for the explanation of the distributions of SOC and TN. SOC and TN contents reduced from northeast to southeast of the CBNC, and the average predicted contents were 13.22 g kg−1 and 1.28 g kg−1, respectively. The total level of C:N ratio was relatively low, suggesting high rates of decomposition. Our study results indicated that the influence of human activities on croplands of the CBNC should not be ignored. It also recommended that the improvement of agricultural management was an approach to re-configure the spatial distributions of SOC and TN in the CBNC of China.

Geospatial modeling approaches for mapping topsoil organic carbon stock in northern part of Mongolia

Soil organic carbon (SOC) is one of the most important indicators of soil quality and agricultural productivity. This paper presents the application of Regression Kriging (RK), geographically weighted regression (GWR) and Geographically Weighted Regression Kriging (GWRK) for prediction of topsoil organic carbon stock in Tarialan. A total of 25 topsoil (0-30 cm) samples were collected from Tarialan soum of Khuvsgul aimag in Mongolia. In this study, seven independent variables were used including normalised difference vegetation index (NDVI), soil adjusted vegetation index (SAVI), normalised difference moisture index (NDMI), land surface temperature (LST) and terrain factors (DEM, Slope, Aspect). We used root-mean-square error (RMSE), mean error (ME) and determination coefficient (R2) to evaluate the performance of these methods. Validation results showed that performance of the GWRK, GWR, and RK approaches were good with not only low values of root-mean-square error (1.38 kg/m2, 1.48 kg/m2, 0.69 kg/m2), mean error (0.28 kg/m2, -0.22 kg/m2, 0.17 kg/m2) but also high values of R2 (0.76, 0.72, 0.94). The estimated SOC stock values ranged from 0.28-16.26 kg/m2, 0.72–15.24 kg/m2, 0.16–15.83 kg/m2 using GWRK, GWR, RK approaches in the study area. The highest average SOC stock value was in the wetland (6.47 kg/m2, 6.08 kg/m2, 6.44 kg/m2) and the lowest was in cropland (1.63 kg/m2, 1.48 kg/m2, 1.80 kg/m2) using these approaches. According to the validation, GWRK, GWR, and RK approaches produced satisfactory results for estimating and mapping SOC stock. However, Regression Kriging was the best model, followed by GWRK and GWR to predict topsoil organic carbon stock in Tarialan.