Dealing with spatial heterogeneity in pointwise-to-gridded- data comparisons

Most studies on validation of satellite trace gas retrievals or atmospheric chemical transport models assume that pointwise measurements, which roughly represent the element of space, should compare well with satellite (model) pixels (grid box). This assumption implies that the field of interest must possess a high degree of spatial homogeneity within the pixels (grid box), which may not hold true for species with short atmospheric lifetimes or in the proximity of plumes. Results of this assumption often lead to a perception of a nonphysical discrepancy between data, resulting from different spatial scales, potentially making the comparisons prone to overinterpretation. Semivariogram is a mathematical expression of spatial variability in discrete data. Modeling the semivariogram behavior permits carrying out spatial optimal linear prediction of a random process field using kriging. Kriging can extract the spatial information (variance) pertaining to a specific scale, which in turn translates pointwise data to a gridded space with quantified uncertainty such that a grid-to-grid comparison can be made. Here, using both theoretical and real-world experiments, we demonstrate that this classical geostatistical approach can be well adapted to solving problems in evaluating model-predicted or satellite-derived atmospheric trace gases. This study suggests that satellite validation procedures using the present method must take kriging variance and satellite spatial response functions into account. We present the comparison of Ozone Monitoring Instrument (OMI) tropospheric NO2 columns against 11 Pandora spectrometer instrument (PSI) systems during the DISCOVER-AQ campaign over Houston. The least-squares fit to the paired data shows a low slope (OMI=0.76×PSI+1.18×1015 molecules cm−2, r2=0.66), which is indicative of varying biases in OMI. This perceived slope, induced by the problem of spatial scale, disappears in the comparison of the convolved kriged PSI and OMI (0.96×PSI+0.66×1015 molecules cm−2, r2=0.72), illustrating that OMI possibly has a constant systematic bias over the area. To avoid gross errors in comparisons made between gridded data vs. pointwise measurements, we argue that the concept of semivariogram (or spatial autocorrelation) should be taken into consideration, particularly if the field exhibits a strong degree of spatial heterogeneity at the scale of satellite and/or model footprints.

A GIS-Based Approach for the Quantitative Assessment of Soil Quality and Sustainable Agriculture

Assessing soil quality is considered one the most important indicators to ensure planned and sustainable use of agricultural lands according to their potential. The current study was carried out to develop a spatial model for the assessment of soil quality, based on four main quality indices, Fertility Index (FI), Physical Index (PI), Chemical Index (CI), and Geomorphologic Index (GI), as well as the Geographic Information System (GIS) and remote sensing data (RS). In addition to the GI, the Normalized Difference Vegetation Index (NDVI) parameter were added to assess soil quality in the study area (western part of Matrouh Governorate, Egypt) as accurately as possible. The study area suffers from a lack of awareness of agriculture practices, and it depends on seasonal rain for cultivation. Thus, it is very important to assess soil quality to deliver valuable data to decision makers and regional governments to find the best ways to improve soil quality and overcome the food security problem. We integrated a Digital Elevation Model (DEM) with Sentinel-2 satellite images to extract landform units of the study area. Forty-eight soil profiles were created to represent identified geomorphic units of the investigated area. We used the model builder function and a geostatistical approach based on ordinary kriging interpolation to map the soil quality index of the study area and categorize it into different classes. The soil quality (SQ) of the study area, classified into four classes (i.e., high quality (SQ2), moderate quality (SQ3), low quality (SQ4), and very low quality (SQ5)), occupied 0.90%, 21.87%, 22.22%, and 49.23% of the total study area, respectively. In addition, 5.74% of the study area was classified as uncultivated area as a reference. The developed soil quality model (DSQM) shows substantial agreement (0.67) with the weighted additive model, according to kappa coefficient statics, and significantly correlated with land capability R2 (0.71). Hence, the model provides a full overview of SQ in the study area and can easily be implemented in similar environments to identify soil quality challenges and fight the negative factors that influence SQ, in addition to achieving environmental sustainability.

Application of Geostatistical Tools to the Geochemical Characterization of the Peloritani Mts (Sicily, Italy) Aquifers

Sources of groundwater contaminants in inhabited areas, located in complex geo-tectonic contexts, are often deeply interlocked, thus, making the discrimination between anthropic and natural origins difficult. In this study, we investigate the Peloritani Mountain aquifers (Sicily, Italy), using the combination of probability plots with concentration contour maps to retrieve an overall view of the groundwater geo-chemistry with a special focus on the flux of heavy metals. In particular, we present a methodology for integrating spatial data with very different levels of precision, acquired before and during the “geomatic era”. Our results depict a complex geochemical layout driven by a geo-puzzle of rocks with very different lithological natures, hydraulically connected by a dense tectonic network that is also responsible for the mixing of deep hydrothermal fluids with the meteoric recharge. Moreover, a double source, geogenic or anthropogenic, was individuated for many chemicals delivered to groundwater bodies. The concentration contour maps, based on the different data groups identified by the probability plots, fit the coherency and congruency criteria with the distribution of both rock matrices and anthropogenic sources for chemicals, indicating the success of our geostatistical approach.

Modelling Spatial Pattern of Anemia and Malnutrition Co-occurrence Among Under-five Children in Ethiopia: A Bayesian Geostatistical Approach

Background: Anemia and Malnutrition among under-five children are one of the major challenges to public health in Ethiopia. While anemia is responsible for delayed child development and growth, malnutrition is associated with the high infant mortality rate in Ethiopia. Method: This study aims to determine the socioeconomic, demographic, and geographical risk factors that simultaneously increase the co-occurrence of anemia and malnutrition among under-five children in Ethiopia. Geostatistical data was obtained from the Ethiopia Demographic and Health Survey for 2011 and 2016. A Bayesian hierarchical linear mixed model was adopted using the stochastic partial differential equation to estimate the spatial pattern of the co-occurrence of anemia and malnutrition in Ethiopia. Result: The findings revealed that gender, maternal education, number of children under five, birth order, preceding birth, contraceptive use, vaccination, marital status, birth weight, diarrhea, and fever are significant risk factors of the co-occurrence of anemia and malnutrition. The findings also reveal the vulnerability of under-five children to the co-occurrence of anemia and malnutrition within the first twenty months after birth and young maternal age. Regarding the geographical aspect, this study found a geographical disparity in the prevalence of anemia and malnutrition in Ethiopia. The highest burden of the co-occurrence of anemia and malnutrition lies in the Northern Gambela, Western Oromia, Northeast Benishangul-gumuz, Central and Northern Amhara, Southern Afar, and parts of Somali. Conclusion: These findings could be utilized by policymakers and intervention programs to simultaneously tackle and contain the prevalence of both anemia and malnutrition. For cost-effective intervention, policies and programs that improve individual-level risk factors of parents and caregivers are a more promising approach to tackle high prevalent regions than the ones on the children and should be considered as an utmost priority in the country.

High-Resolution North Sulawesi Drought Hazzard Mapping Based on Consecutive Dry Days (CDD)

North Sulawesi is one of the Province in northern Indonesia with high spatial annual rainfall variations and influenced by global climate anomaly that can lead to extreme events and disaster occurrence, such as flood, landslide, drought, etc. The purpose of this study is to generate high-resolution meteorological hazard map based on long-term historical consecutive dry days (CDD) over the North Sulawesi region. CDD was calculated based on observed daily precipitation data from Indonesia Agency for Meteorology, Climatology, and Geophysics (BMKG) surface observation station network (CDDobs) and the daily-improved Climate Hazards group Infrared Precipitation with Stations (CHIRPS) version 2.0 (CDDCHIRPS) during 1981 – 2010 period. The Japanese 55-year Reanalysis (JRA-55) data obtained from iTacs (Interactive Tool for Analysis of the Climate System) with the same time scale period also used to explain physical – dynamical atmospheric properties related to drought hazard over this region. The Geostatistical approach using regression kriging method was applied as spatial interpolation technique to generate high resolution gridded (0.05° × 0.05°) drought hazard map. This method combines a regression of CDDobs as dependent variable (target variable) on CDDCHIRPS as predictors with kriging of the prediction residuals. The results show that most of the areas were categorized as medium drought hazard level with CDD values ranging from 80-100 days. Meanwhile, small islands around main Sulawesi island such as Sangihe and Karakelong island are dominated by low drought hazard levels with CDD values ranging from 50-60 days. The highest levels of drought hazard area are located in South Bolaang Mongondow Regency.

Photoneutron activation analysis and multivariate geostatistical approach to assess the spatial occurrence of trace heavy metals in Crisuri Basin

The Crisuri Basin, with its total length of 1093 km (among which 670 km in Romania), is subjected to an effective environmental monitoring system consisting of 18 water stations network. Their integrated information is used for the water assessment problem to understand the biogeochemical processes leading to significant pollution levels in some running water sectors. The goal of this research is to predict the most important factors causing the change of the geochemical measured parameters of some components of the Crisuri water resources. In the present paper, we developed a multivariate statistical model to estimate the spatiotemporal distribution of heavy metals in the field and to identify the contamination sources of Basin River waters. Two methods were deployed as an overall approach to fulfill the proposed objectives, namely the photoneutron activation analysis followed by high-resolution gamma-ray spectrometry and the multivariate statistical analysis. The elements analyzed by different analytical techniques and introduced in databases were As, Cd, Ca, Cu, Fe, Mg, Hg, Na, Ni, Pb, Zn, N-NH4, N-NO2, N-NO3, P-PO4, fixed residues, S-SO4, Cl, phenols and, additional oil compounds. By combining the spatially distributed geochemical data on trace heavy metals with the spatially distributed geophysical data, we obtained the most significant fingerprint factors and their associated uncertainty information concerning the water quality.

Dealing with Spatial Heterogeneity in Pointwise to Gridded Data Comparisons

Atmospheric modelers and the trace gas retrieval community typically presuppose that pointwise measurements, which roughly represent the element of space, should compare well with satellite (model) pixels (grids). This assumption implies that the field of interest must possess a high degree of spatial homogeneity within the pixels (grids), which may not hold true for species with short atmospheric lifetimes or in the proximity of plumes. Results of this assumption often lead to a perception of a nonphysical discrepancy between data, resulting from different spatial scales, potentially making the comparisons prone to overinterpretation. Semivariogram is a mathematical expression of spatial variability in discrete data. Modeling the semivariogram behavior permits carrying out spatial optimal linear prediction of a random process field using kriging. Kriging can extract the spatial information (variance) pertaining to a specific scale, which in turn translating pointwise data to a grid space with quantified uncertainty such that a grid-to-grid comparison can be made. Here, using both theoretical and real-world experiments, we demonstrate that this classical geostatistical approach can be well adapted to solving problems in evaluating model-predicted or satellite-derived atmospheric trace gases. This study demonstrates that satellite validation procedures must take kriging variance and satellite spatial response functions into account. We present the comparison of Ozone Monitoring Instrument (OMI) tropospheric NO2 columns against 11 Pandora Spectrometer Instrument (PSI) systems during the DISCOVER-AQ campaign over Houston. The least-squares fit to the paired data shows a low slope (OMI=0.76×PSI+1.18×1015 molecules cm−2, r2=0.67) which is indicative of varying biases in OMI. This perceived slope, induced by the problem of spatial scale, disappears in the comparison of the convolved kriged PSI and OMI (0.96×PSI+0.66×1015 molecules cm−2, r2=0.72) illustrating that OMI possibly has a constant systematic bias over the area. To avoid gross errors in comparisons made between gridded data versus pointwise measurements, we argue that the concept of semivariogram (or spatial auto-correlation) should be taken into consideration, particularly if the field exhibits a strong degree of spatial heterogeneity at the scale of satellite and/or model footprints.

Spatial distribution of sediment archaeal and bacterial communities relates to the source of organic matter and hypoxia – a biogeographical study on Lake Remoray (France)

Bottom waters hypoxia spreads in many lakes worldwide causing severe consequences on whole lakes trophic network. Here, we aimed at understanding the origin of organic matter stored in the sediment compartment and the related diversity of sediment microbial communities in a lake with deoxygenated deep water layers. We used a geostatistical approach to map and compare both the variation of organic matter and microbial communities in sediment. Spatialisation of C/N ratio and δ13C signature of sediment organic matter suggested that Lake Remoray was characterized by an algal overproduction which could be related to an excess of nutrient due to the close lake-watershed connectivity. Three spatial patterns were observed for sediment microbial communities after the hypoxic event, each characterized by specific genetic structure, microbial diversity and composition. The relative abundance variation of dominant microbial groups across Lake Remoray such as Cyanobacteria, Gammaproteobacteria, Deltaproteobacteria and Chloroflexi provided us important information on the lake areas where hypoxia occurs. The presence of methanogenic species in the deeper part of the lake suggests important methane production during hypoxia period. Taken together, our results provide an extensive picture of microbial communities' distribution related to quantity and quality of organic matter in a seasonally hypoxic lake.

Weekly Mapping of Sea Ice Freeboard in the Ross Sea from ICESat-2

NASA’s ICESat-2 has been providing sea ice freeboard measurements across the polar regions since October 2018. In spite of the outstanding spatial resolution and precision of ICESat-2, the spatial sparsity of the data can be a critical issue for sea ice monitoring. This study employs a geostatistical approach (i.e., ordinary kriging) to characterize the spatial autocorrelation of the ICESat-2 freeboard measurements (ATL10) to estimate weekly freeboard variations in 2019 for the entire Ross Sea area, including where ICESat-2 tracks are not directly available. Three variogram models (exponential, Gaussian, and spherical) are compared in this study. According to the cross-validation results, the kriging-estimated freeboards show correlation coefficients of 0.56–0.57, root mean square error (RMSE) of ~0.12 m, and mean absolute error (MAE) of ~0.07 m with the actual ATL10 freeboard measurements. In addition, the estimated errors of the kriging interpolation are low in autumn and high in winter to spring, and low in southern regions and high in northern regions of the Ross Sea. The effective ranges of the variograms are 5–10 km and the results from the three variogram models do not show significant differences with each other. The southwest (SW) sector of the Ross Sea shows low and consistent freeboard over the entire year because of the frequent opening of wide polynya areas generating new ice in this sector. However, the southeast (SE) sector shows large variations in freeboard, which demonstrates the advection of thick multiyear ice from the Amundsen Sea into the Ross Sea. Thus, this kriging-based interpolation of ICESat-2 freeboard can be used in the future to estimate accurate sea ice production over the Ross Sea by incorporating other remote sensing data.