A Categorical Machine Learning Approach to Predicting Areas of Shallow Coastal Groundwater.

Extensive and unregulated groundwater extraction for irrigation in the arid inland basins of Northwest China has led to a continuous increase in groundwater depth in agricultural irrigation areas. This has significantly altered the distribution of soil ions, making it difficult to predict their evolution and dynamic patterns. In this study, we used a space-for-time substitution approach to elucidate the evolution of the soil ion distribution under changing groundwater depths. Experiments were conducted in three typical irrigation areas with varying groundwater depths, that is, below 5 m, 5–10 m, and above 10 m in Korla, Xinjiang, China. Soil samples were collected from five profiles at depths of 0–180 cm to measure the soil moisture, salinity, and major ion content. An innovative research framework was developed to examine the relationship between groundwater depth and soil ion distribution using ion ratios, principal components, hierarchical clustering, and correlation analyses. This framework aims to reveal the dynamics, correlations, and mechanisms of soil moisture, salinity, ion distribution, and representative ion composition as groundwater depth increases in the arid agricultural irrigation areas of Northwest China. The results showed that as groundwater depth increased, the soil chemical type shifted from Ca-SO4 to Na-SO4 and mixed types, with an increase in SO42− and Na+ content in the soil profile. Soil moisture, salinity, sodium adsorption ratio (SAR), and total dissolved solids (TDS) were significantly higher in shallow groundwater than in deep groundwater. Groundwater depth was negatively correlated with soil moisture, salinity, and major cations and anions (K+, Na+, Ca2+, Mg2+, Cl−, SO42−, and NO3−). Meanwhile, a positive correlation exists between groundwater depth and CO32−. The dynamic distribution of soil ions is primarily governed by groundwater depth and is influenced by multiple factors. Evaporation is the dominant factor in shallow groundwater areas, whereas the mineral composition of rocks plays a crucial role in deep groundwater areas. These findings provide scientific support for strategic agricultural water-resource management policies and sustainable development strategies in arid regions.

Groundwater is an important ecological water source in arid areas. Groundwater depth (GWD) is an important indicator that affects vegetation growth and soil salinization. Clarifying the coupling relationship between vegetation, groundwater, and soil in arid areas is beneficial to the prevention of environmental problems such as desertification and salinization. Existing studies lack research on the water–soil–vegetation relationship in typical areas, especially in shallow groundwater areas. In this study, the shallow groundwater area in Minqin, northwest China, was taken as study area, and vegetation surveys and soil samples collection were conducted. The relationships between vegetation fractional coverage (VFC) and GWD, soil salinity, soil moisture, and precipitation were comprehensively analyzed. The results showed low soil salinity in the riparian zone and high soil salinity in other shallow-buried areas with salinization problems. Soil salinity was negatively correlated with VFC (R = −0.4). When soil salinity >3 g/kg, VFC was less than 20%. Meanwhile, when GWD >10 m, VFC was usually less than 15%. In the areas with soil salinity <3 g/kg, when GWD was in the range of 4–10 m, VFC was positively correlated with soil moisture content (R = 0.99), and vegetation growth mainly depended on surface soil water, which was significantly affected by precipitation. When GWD was less than 4 m, VFC was negatively correlated with GWD (R = −0.78), and vegetation growth mainly relied on groundwater and soil water. There are obvious ecological differences in the shallow-buried areas in Minqin. Hence, it is reasonable to consider zoning and grading policies for ecological protection.

Bioretention cells, which are generally effective in controlling surface runoff and recharging groundwater, have been widely adopted as low impact development practices. However, shallow groundwater has limited their implementation in some locations due to the potential problems of a reduction in surface runoff control, groundwater pollution, and continuous groundwater drainage through the underdrain. Many guidelines have established minimum requirements for the groundwater depth below bioretention cells, but they may not be optimized for certain environmental conditions and bioretention cell designs. This study made use of a variably saturated flow model to examine the hydrologic performance of a single bioretention cell in shallow groundwater with event‐based simulations, considering a wide range of initial groundwater depths, media and in situ soil types, surface runoff loads, and underdrain sizes. Performance indicators (e.g., runoff reduction, time for infiltrated water to reach the bioretention cell bottom and the groundwater table, and height and dissipation time of groundwater mound) were evaluated to examine the processes of runoff generation, the formation and dissipation of groundwater mounds, and the bioretention cell's performance in a shallow groundwater environment. The most influential factors were the initial groundwater depth, the hydraulic conductivity of the media soil, and the rainfall runoff load. With a deeper initial groundwater table, infiltrated water took longer to reach the bioretention cell bottom and groundwater table. Groundwater mounds, however, took longer to dissipate even though they were smaller. The groundwater quality can be better protected if relatively less‐permeable soil types (e.g., sandy loam) are used as the media, although it may compromise the performance in runoff quantity control. However, only very high surface runoff loads would cause concerns regarding a reduction in runoff quantity control and possible groundwater contamination due to the shallow groundwater. A distance of 1.5–3 m between the bioretention cell bottom and the groundwater table is generally sufficient. The results of this study could help to guide the planning and design of bioretention cells in areas of shallow groundwater.

Due to the large spatial variation of groundwater depth, it is very difficult to determine suitable irrigation schedules for crops in shallow groundwater area. A zoning optimization method of irrigation schedule is proposed here, which can solve the problem of the connection between suitable irrigation schedules and different groundwater depths in shallow groundwater areas. The main results include: (1) Taking the annual mean groundwater depth 2.5 m as the dividing line, the shallow groundwater areas were categorized into two irrigation schedule zones. (2) On the principle of maximizing the yield, the optimized irrigation schedule for spring wheat in each zone was obtained. When the groundwater depth was greater than 2.5 m, two rounds of irrigation were chosen at the tillering–shooting stage and the shooting–heading stage with the irrigation quota at 300 mm. When the groundwater depth was less than 2.5 m, two rounds of irrigation were chosen at the tillering–shooting stage, and one round at the shooting–heading stage, with the irrigation quota at 240 mm. The main water-saving effect of the optimized irrigation schedule is that the yield, the soil water use rate, and the water use productivity increased, while the irrigation amount and the ineffective seepage decreased.

In the saline water area of our research, deep groundwater was over-pumped for agricultural irrigation which resulted in a decline of the deep groundwater level and an increase in the shallow groundwater table. Soil salination was also aggravated due to the strong evapotranspiration (ET) in the shallow groundwater areas, where ET removes water vapor from the unsaturated zone (ETu), and the groundwater (ETg). Joint utilities of multiple water sources of transferred water and local shallow and deep groundwater are essential for reasonable management of irrigation water. However, it is still difficult to distinguish ETu and ETg in coupled management of unsaturated zone and groundwater, which account for the water balance in utilities of multiple water sources in a regional scale. In this paper, we used an RS-based ETWatch model as a source of evapotranspiration data coupled with UZF-MODFLOW, an integrated hydrological model of the unsaturated–saturated zone, to estimate the ETg and ETu on a regional scale. It was shown that the coupled model (ETWatch-UZF-MODFLOW) avoids the influence of ETu on the groundwater balance calculation and improves the accuracy of the groundwater model. The model was used in the simulation and prediction of groundwater level. The eastern North China Plain (NCP) was selected as the study area where shallow groundwater was saline water and deep groundwater cone existed. We compared four different scenarios of irrigation methods, including current irrigation scenario, use of saline water, limited deep groundwater pumping, use of multiple water sources of transferred water and local groundwater. Results indicate that the total ETg for the four scenarios in the study area from 2013 to 2030 is 119 × 108 m3, 81.9 × 108 m3, 85.0 × 108 m3, and 92.3 × 108 m3, respectively, and the proportion of ETg to total ET was 6.85%, 4.79%, 4.97%, 5.37%. However, in regions where the groundwater depth is less than 3 m, ETg accounts for 12% of the total ET, indicating that groundwater was one of the main sources of evapotranspiration in shallow groundwater depth area.

As drought variability increases in forests around the globe, it is critical to evaluate and understand ecosystem attributes that ameliorate drought impacts. Trees in arid and semi‐arid ecosystems can sustain tree growth and transpiration during drought by accessing shallow groundwater, yet the extent to which groundwater influences forest growth and transpiration in humid environments has largely been unexplored. We quantified groundwater's influence on tree growth and transpiration in northern humid forests with sandy soils. We hypothesized that even in wet regions, soil droughts occur relatively frequently in forests with sandy soils and result in water stress and reduced tree growth. Further, we hypothesized these reductions in productivity are ameliorated if the forest can access shallow groundwater during dry conditions. We evaluated tree growth responses using tree cores in Pinus resinosa trees and estimated forest groundwater use from diel water table fluctuations across sites covering a 1‐ to 9‐m depth‐to‐groundwater (DTG) gradient. In areas of shallow groundwater (DTG < 2.5 m), we observed twice as much tree growth and high, frequent groundwater use (up to 81% of non‐rainy summer days). Groundwater's influence on tree growth and transpiration declined as groundwater deepened along the DTG gradient in the range 1–5 m below land surface. These findings suggest that water provided by a shallow water table subsidizes evapotranspiration in humid forests and results in enhanced tree growth. Our research provides a basis for understanding the role of groundwater in conferring drought resistance in humid forests to help guide sustainable water and forest management decisions.

The random forest is stochastic a forest establishment, there are many decision trees; there is no correlation between each decision tree random forest. The establishment of each decision tree, using the random sampling process is put back, and then uses the voting form of classification and prediction. The algorithm can solve the bottleneck in the performance of a single classifier, so it is widely used in many aspects. Of course, the algorithm also has some room for improvement, according to the random forest algorithm to deal with unbalanced data set when running low efficiency, this paper puts forward approaches to the problem are not a new balance at the same time as the calculation process, showing the growth of the index value, how to improve the prediction speed and shorten the running time, according to the characteristics of the random forest algorithm in the construction process is put forward Based on the domestic and foreign literatures, this paper mainly studies the optimization of random forest from two aspects. Random forest algorithm is an ensemble learning method in the field of machine learning. It is integrated with the classification results of multiple decision trees to form a global classifier. The random forest algorithm compared with other classification algorithms have many advantages, the classified effect advantage is reflected in the classification accuracy and the generalization error is small and has the ability to deal with high dimensional data, the training process of the advantages of learning algorithm of quick and easy parallelization. Based on these two advantages, random forest algorithm has been widely used, and it has become one of the priorities to deal with classification problem. However, when the data type of the unbalanced distribution of the situation, that is the number one category of samples is far less than other types of samples, random forest algorithm will appear ineffective, the generalization error of variable classification problem and a series of. So far, there is not much research on the problem of unbalanced data for random forest classification, and there is no direct and effective method. Some just combine the general processing methods of unbalanced data, such as sampling technique or cost sensitive method. So it is a significant research problem to improve the classification effect of unbalanced data from the random forest algorithm level. Based on this research, this paper analyzes the key steps in the analysis of the effect of random forest classification, and designs a solution to deal with unbalanced data. In this paper, we propose an improved random forest algorithm to deal with the problem of imbalanced data classification by studying the classification method of unbalanced data and the random forest algorithm. Mainly from two aspects of the sub space selection and model integration of random forest. In this paper, the influence of the balanced sampling on the algorithm is also combined with the experimental results. Finally, verify the improved random forest algorithm in unbalanced classification results on public data sets, compared to the original random forest algorithm, in most indicators (cross validation accuracy, AUC index, Kappa coefficient and F1-Measure index) have obvious improvement. The importance of subspace selection and model optimization for random forest algorithm is demonstrated. The research content of this paper has an important academic significance and practical value to guide the classification of imbalanced data, and can be applied to the field of spam detection, anomaly detection, medical diagnosis, DNA sequence identification and so on.

BackgroundThe random forests algorithm is a type of classifier with prominent universality, a wide application range, and robustness for avoiding overfitting. But there are still some drawbacks to random forests. Therefore, to improve the performance of random forests, this paper seeks to improve imbalanced data processing, feature selection and parameter optimization.ResultsWe propose the CURE-SMOTE algorithm for the imbalanced data classification problem. Experiments on imbalanced UCI data reveal that the combination of Clustering Using Representatives (CURE) enhances the original synthetic minority oversampling technique (SMOTE) algorithms effectively compared with the classification results on the original data using random sampling, Borderline-SMOTE1, safe-level SMOTE, C-SMOTE, and k-means-SMOTE. Additionally, the hybrid RF (random forests) algorithm has been proposed for feature selection and parameter optimization, which uses the minimum out of bag (OOB) data error as its objective function. Simulation results on binary and higher-dimensional data indicate that the proposed hybrid RF algorithms, hybrid genetic-random forests algorithm, hybrid particle swarm-random forests algorithm and hybrid fish swarm-random forests algorithm can achieve the minimum OOB error and show the best generalization ability.ConclusionThe training set produced from the proposed CURE-SMOTE algorithm is closer to the original data distribution because it contains minimal noise. Thus, better classification results are produced from this feasible and effective algorithm. Moreover, the hybrid algorithm's F-value, G-mean, AUC and OOB scores demonstrate that they surpass the performance of the original RF algorithm. Hence, this hybrid algorithm provides a new way to perform feature selection and parameter optimization.

The Xingtai piedmont plain in Hebei Province is a representative area in northern China where endemic fluorosis is serious and shallow high-F groundwater is distributed. The fluoride concentration of the shallow high-F groundwater is usually in the range of 1.0 — 2.0mg/l in this area. The F-bearing minerals for biotite and hornblende and the adsorbed and dispersed fluorides are distributed in the Quaternary deposits, particularly the Holocene deposits are the main supply source of fluorine for the shallow groundwater. In the distribution area of the shallow high-F groundwater, the unsaturated soils and groundwater constitute an interactive hydrogeochemical system. The unsaturated soils have an important effect on fluoride migration and accumulation. According to the experimental analysis and the coupled model for fluoride hydrodynamic transport-chemical reaction, the increase of F concentration of the shallow high-F groundwater can be effictively controlled by regulating the burial depth of the shallow groundwater to be greater than 4.0–6.0m. In the shallow ground water F−, MgF+ and CaF+ are the major fluorine species, and the ratioes of aMgF +, and aCaF + to the molalities of total fluoride are 79.35–96.51%, 3.09–19.17% and 0.32–3.01%, respectively. These F species have different influences on endemic fluorosis, of which MgF+ may have more important effect on endemic fluorosis in the area. The mass-balance model of fluoride suggests that average fluoride contents of 1.55mg/l in the shallow high-fluoride ground water are mainly originated from the dissolution of 0.0393mmol/KgH2O fluorine-bearing biotite along flow path from the mountain to the plain.

Landslide susceptibility map refinement using PSInSAR data

The Yudong Plain is in the eastern part of Henan Province, China, where there is little rain and high evaporation. Compared to other areas in Henan Province, the groundwater fluorine content is generally high, which affects the health of residents. Based on the systematic analysis of water chemistry data of shallow and mid-depth groundwater samples in the Yudong Plain, the causes of shallow and mid-depth high-fluorine groundwater in the Yudong Plain were explored using mathematical statistics, spatial interpolation, and ion ratios. The results show that the fluorine contents of both shallow and mid-depth groundwater in the study area are high. The shallow samples had fluorine contents ranging from 0.1 to 4.89mg/L, with an exceedance rate of 48% and an average content of 1.15mg/L. The fluorine content of mid-depth samples ranged from 0.14 to 3.32mg/L, with an exceedance rate of 68% and an average content of 1.33mg/L. The shallow high-fluorine groundwater is mainly distributed in the central low-lying area, and its main hydrochemical type is HCO3-Na·Mg; the mid-depth high-fluorine groundwater is mainly distributed in strips in the north and east of the study area, and its main water chemistry type is HCO3-Na. Fluorine enrichment in shallow groundwater in the study area is controlled by rock weathering, evaporation concentration, and competitive adsorption, while leaching and dissolution of fluorine-containing minerals in sedimentary strata are the main factors influencing fluorine enrichment in mid-depth groundwater. The results of the human health risk assessment (HRA) showed that the mean non-carcinogenic hazard quotients (HQs) in shallow groundwater were 0.95, 0.64, 0.57, and 0.55 for infants, children, teenagers, and adults, respectively, while the mean non-carcinogenic HQs in mid-depth groundwater were 1.11, 0.74, 0.66, and 0.63, respectively. The study provides a scientific basis for the rational development and use of groundwater in the area and offers theoretical support for the prevention and control of groundwater pollution.

ABSTRACTWe evaluated invasive exotic weeds and plant species diversity in relation to depth to groundwater (DTW) and microtopography in areas with DTW from 0·3–4 m in Owens Valley, California. Transects dominated by common plant species of the area were read at 1‐cm intervals, and species cover was obtained at different scales: 1‐m transect portions (microsites), whole transects (68 m average length), and the whole study area. Species richness and the Shannon–Wiener diversity index were obtained in microsites. DTW and microtopographical variation (soil‐surface relative elevations) were jointly measured along the vegetation transects. Found in 34% of the sampled transects, the annual Bassia hyssopifolia was the most common exotic weed of the area. Its cover in the whole study area was only 0·9%, but it was the second most dominant species (19% cover) in microsites. B. hyssopifolia coexisted mainly with herbaceous species, typically in places with shallow groundwater (average <1·5 m) and microtopographical depressions. However, B. hyssopifolia cover was not affected by variations in DTW or relative elevation. In contrast, B. hyssopifolia cover was negatively associated with perennial cover and species richness. Species diversity (Shannon–Wiener) slightly decreased with the increase in DTW but not with changes in microtopography. In contrast, species richness clearly varied with the identity of the dominant species in microsites. Our study suggests that the variation in community‐intrinsic factors such as competition and diversity may play a greater role in the growth of invasive plants than the variation in physical factors such as groundwater and microtopography. Copyright © 2011 John Wiley & Sons, Ltd.

Soil nutrients are vital for plant growth and survival and present a crucial role in terrestrial function and productivity. However, little is known about the effect mechanism of groundwater table on soil nutrients in an arid desert ecological system. This study investigated the impacts of groundwater depth on the concentrations of soil organic carbon (C), available nitrogen (N), phosphorus (P), and potassium (K) at shallow groundwater depths (0.4, 0.8, 1.2, 1.8, and 2.2 m) and field deep groundwater depths (2.5, 4.5, and 11.0 m) in a desert-oasis ecotone in Central Asia in 2015 and 2016. Soil nitrate-N, inorganic-N, soil available P, and K concentrations were significantly affected by shallow and field deep groundwater. Groundwater depths did not alter soil ammonium-N concentration. Soil organic C concentration was influenced by field deep groundwater depth. Structural equation model showed that groundwater depth directly affected soil nitrate-N and K concentrations and indirectly altered the soil inorganic-N, soil organic C and available P concentrations in shallow groundwater. Moreover, groundwater depth directly influenced soil nitrate-N and soil organic C, available P and K concentrations and indirectly affected soil inorganic-N concentration in deep groundwater. Hence, groundwater depth should be considered one of the most critical environmental factors affecting soil nutrient variation in an arid desert. This study provides new insights into the soil nutrient variation under a declining groundwater depth in a hyper-arid ecosystem.

The depth of shallow groundwater significantly influences crop growth and yield by altering the physicochemical properties of farmland soil profiles. Concurrently, shallow groundwater is subject to various changes, and it remains unclear how alterations in shallow groundwater depth within field soil impact soil physicochemical properties, crop yields, and the overall dynamics of groundwater transformations. To address these uncertainties, this study utilized a sample plot equipped with a volume lysimeter and implemented four distinct groundwater depths as treatment conditions: G0 (no groundwater depth), G1 (a groundwater depth of 40 cm), G2 (a groundwater depth of 70 cm), G3 (a groundwater depth of 110 cm), and G4 (a groundwater depth of 150 cm). This study was carried out on a weekly basis to monitor fluctuations in ion content in shallow groundwater and soil moisture after the summer maize harvest, and special attention was afforded to non-irrigation conditions. This study also scrutinized the distribution of salt and nutrients in soil profiles and assessed changes in summer maize yield. Very interesting findings were obtained by conducting the study. Firstly, the shallower the groundwater depth, the higher the water and salt content of the soil surface. Small, frequent rainfall events (precipitation ≤ 25 mm) facilitated the effective removal of salt from the soil surface. Despite increased rainfall contributing to salt ion dilution in groundwater, the risk of soil surface salinization increased at the surface level. Secondly, a linear relationship existed between groundwater depth and surface soil moisture and salt content. With every 10 cm increase in groundwater depth, the surface soil moisture and salt content decreased by 0.56% and 0.06 g/kg, respectively. Soil nutrients tended to accumulate in the surface layer, with nutrient content increasing with depth. However, C/N was not notably affected by groundwater depth. Thirdly, Na+ and K+ consistently dominated the soil surface. As soil salinity increased, the prevalence of Cl− and SO42− increased, with the rate of SO42− increase surpassing that of chlorine. HCO3− altered by rainfall served as an indicator of soil alkalization characteristics, while Na+ and K+ in soil, along with Cl− and SO42− derived from groundwater, represented soil salt composition and salinization trends. Ultimately, under the conditions of this study, the most favorable groundwater depth for the growth of summer maize was determined to be 1.1 m. Analyzing the impact of different shallow groundwater depths on the physicochemical properties of farmland soil enhances our understanding of the mechanisms of interaction between groundwater and soil in agricultural ecosystems. This knowledge is instrumental in significantly improving the soil environment, thereby ensuring optimal crop yields.

The effect of groundwater nutrient inputs on river and estuary water quality and the potential impacts of urbanization on groundwater are central concerns in many coastal areas. It has been previously identified that dissolved organic nitrogen (DON) can be the dominant form of total dissolved nitrogen (TDN) in some aquifers. However, there is a paucity of evidence about the sources and flow paths of DON, relative to inorganic nitrogen in groundwater. DON and dissolved organic carbon/DON were first compared against different landscape variables in this study, and no significant relationships were found. However, the relationships became statistically significant when shallow samples (sampling depth < 10 m) were separated from deep samples. A random forest model and sensitivity analysis were then applied to further our understanding of the ecohydrological drivers and seasonal patterns that shape DON variability. The random forest algorithm was built to classify 171 groundwater wellbores into three classes (low: <0.5 mg/L; medium: 0.5–2.5 mg/L; and high: >2.5 mg/L) which achieved 72% classification accuracy using landscape characteristics, hydrological conditions, and temporal information. The results indicated that the effects of landscapes on sandy shallow groundwater DON were controlled both by certain landscape characteristics and depth to groundwater. A conceptual model of groundwater DON is therefore proposed where the balance of exposure and processing time scales from the surface to groundwater is the critical control on the preservation of landscape signatures; we expect that this conceptual model would be applicable for other sandy, shallow groundwater areas.