Groundwater hydrochemical signatures, nitrate sources, and potential health risks in a typical karst catchment of North China using hydrochemistry and multiple stable isotopes

Nitrate pollution in aquatic ecosystems has received growing concern, particularly in fragile karst basins. In this study, hydrochemical compositions, multiple stable isotopes (δ2H-H2O, δ18Ο-Η2Ο, δ15Ν-ΝΟ3-, and δ18Ο-ΝΟ3-), and Bayesian stable isotope mixing model (MixSIAR) were applied to elucidate nitrate pollution sources in groundwater of the Yangzhuang Basin. The Durov diagram identified the dominant groundwater chemical face as Ca-HCO3 type. The NO3- concentration ranged from 10.89 to 90.45 mg/L (average 47.34 mg/L), showing an increasing trend from the upstream forest and grassland to the downstream agricultural dominant area. It is worth noting that 47.2% of groundwater samples exceeded the NO3- threshold value of 50 mg/L for drinking water recommended by the World Health Organization. The relationship between NO3-/Cl- and Cl- ratios suggested that most groundwater samples were located in nitrate mixed endmember from agricultural input, soil organic nitrogen, and manure & sewage. The Self-Organizing Map (SOM) and Pearson correlations analysis further indicated that the application of calcium fertilizer, sodium fertilizer, and livestock and poultry excrement in farmland elevated NO3- level in groundwater. The output results of the MixSIAR model showed that the primary sources of NO3- in groundwater were soil organic nitrogen (55.3%), followed by chemical fertilizers (28.5%), sewage & manure (12.7%), and atmospheric deposition (3.4%). Microbial nitrification was a dominant nitrogen conversion pathway elevating NO3- levels in groundwater, while the denitrification can be neglectable across the study area. The human health risk assessment (HHRA) model identified that about 88.9%, 77.8%, 72.2%, and 50.0% of groundwater samples posing nitrate's non-carcinogenic health hazards (HQ > 1) through oral intake for infants, children, females, and males, respectively. The findings of this study can offer useful biogeochemical information on nitrogen pollution in karst groundwater to support sustainable groundwater management in similar human-affected karst regions.

Tracing nitrate origins and transformation processes in groundwater of the Hohhot Basin's Piedmont strong runoff zone through dual isotopes and hydro-chemical analysis

Nitrate, which poses a serious threat to the drinking water supply, is one of the most prevalent anthropogenic groundwater contaminants worldwide. With the development of the chemical industry, the nitrate pollution of groundwater in the Piedmont strong runoff zone of the Hohhot Basin, which is the main groundwater extraction area, is becoming increasingly severe. The special hydrogeological and complex pollution conditions in the study area make it difficult to identify nitrate sources and transformation processes. In order to identify the results more accurately, this study combined water chemistry, multivariate statistical analysis and isotope tracer methods to determine the sources and transformation processes of nitrate in the study area. The results showed that the groundwater in the eastern part of the study area (ESA) was clearly affected by anthropogenic activities, and its nitrate was mainly from nitrification of ammonia in industrial wastewater, nitrate in industrial wastewater (the sum of the two contributions was 62.2 %), and nitrate in manure (20.5 %). The hydrogeochemical characteristics of groundwater in the western part of the study area (WSA) are the same as those of natural groundwater in the Piedmont strong-runoff zone. The nitrate in groundwater in the WSA was mainly derived from soil nitrogen (63.8 %) and ammonia fertilizer (28.8 %). Nitrification and denitrification occurred only locally in the aquifer of the study area and were more pronounced in the ESA. Meanwhile, the transformation processes of nitrate in groundwater in the ESA and WSA was significantly influenced by contamination with chlorinated hydrocarbon volatile organic compounds and hydrogeological conditions, respectively. These findings provide a scientific basis for the development of groundwater pollution prevention measures in the study area and guide the traceability of nitrate in groundwater in areas with similar hydrogeological and pollution conditions.

Riverine nitrate source identification combining δ15N/δ18O-NO3- with Δ17O-NO3- and a nitrification 15N-enrichment factor in a drinking water source region

Dual nitrate isotopes (δ15N/δ18O-NO3-) are an effective tool for tracing nitrate sources in freshwater systems worldwide. However, the initial δ15N/δ18O values of different nitrate sources might be altered by isotopic fractionation during nitrification, thereby limiting the efficiency of source apportionment results. This study integrated hydrochemical parameters, site-specific isotopic compositions of potential nitrate sources, multiple stable isotopes (δD/δ18O-H2O, δ15N/δ18O-NO3- and Δ17O-NO3-), soil incubation experiments assessing the nitrification 15N-enrichment factor (εN), and a Bayesian mixing model (MixSIAR) to reduce/eliminate the influence of 15N/18O-fractionations on nitrate source apportionment. Surface water samples from a typical drinking water source region were collected quarterly (June 2021 to March 2022). Nitrate concentrations ranged from 0.35 to 3.06 mg/L (mean = 0.78 ± 0.46 mg/L), constituting ∼70 % of total nitrogen. A MixSIAR model was developed based on δ15N/δ18O-NO3- values of surface waters and the incorporation of a nitrification εN (-6.9 ± 1.8 ‰). Model source apportionment followed: manure/sewage (46.2 ± 10.7 %) > soil organic nitrogen (32.3 ± 18.5 %) > nitrogen fertilizer (19.7 ± 13.1 %) > atmospheric deposition (1.8 ± 1.6 %). An additional MixSIAR model coupling δ15N/δ18O-NO3- with Δ17O-NO3- and εN was constructed to estimate the potential nitrate source contributions for the June 2021 water samples. Results revealed similar nitrate source contributions (manure/sewage = 43.4 ± 14.1 %, soil organic nitrogen = 29.3 ± 19.4 %, nitrogen fertilizer = 19.8 ± 13.8 %, atmospheric deposition = 7.5 ± 1.6 %) to the original MixSIAR model based on εN and δ15N/δ18O-NO3-. Finally, an uncertainty analysis indicated the MixSIAR model coupling δ15N/δ18O-NO3- with Δ17O-NO3- and εN performed better as it generated lower uncertainties with uncertainty index (UI90) of 0.435 compared with the MixSIAR model based on δ15N/δ18O-NO3- (UI90 = 0.522) and the MixSIAR model based on δ15N/δ18O-NO3- and εN (UI90 = 0.442).

Determination of the water-use patterns for two xerophyte shrubs by hydrogen isotope offset correction

The stable hydrogen and oxygen isotope technique is typically used to explore plant water uptake; however, the accuracy of the technique has been challenged by hydrogen isotope offsets between plant xylem water and its potential source water. In this study, the soil hydrogen and oxygen isotope waterline was used to correct the hydrogen isotope offsets for Salix psammophila and Caragana korshinskii, two typical shrub species on the Chinese Loess Plateau. Five different types of isotopic data [(i) δ18O, (ii) δ2H, (iii) combination δ18O with δ2H, (iv) corrected δ2H and (v) combination δ18O with corrected δ2H] were separately used to determine the water-use patterns of the two shrubs. The δ2H offset values of S. psammophila and C. korshinskii did not show significant temporal variation among the sampling months (May, July and September) but showed notable differences between the two shrubs (-0.4 ± 0.5‰ in S. psammophila vs -4.3 ± 0.9‰ in C. korshinskii). The obtained water absorption proportion (WAP) of S. psammophila in the different soil layers (0-20, 20-60 and 60-200 cm) did not differ significantly among the five different input data types. However, compared with the input data types (iii) and (v), the data types (i), (ii) and (iv) overestimated the WAP of C. korshinskii in the 0-20 cm soil layer and underestimated that in the 60-200 cm layer. The data type (iii) overestimated the WAP of C. korshinskii in 0-20 cm soil layer (25.9 ± 0.8%) in July in comparison with the WAP calculated based on data type (v) (19.1 ± 1.1%). The combination of δ18O and corrected δ2H, i.e., data type (v), was identified as the best data type to determine the water use patterns of C. korshinskii due to the strong correlation between the calculated WAP and soil water content and soil sand content. In general, S. psammophila mainly used (57.9-62.1%) shallow soil water (0-60 cm), whereas C. korshinskii mainly absorbed (52.7-63.5%) deep soil water (60-200 cm). We confirm that the hydrogen isotope offsets can cause significant errors in determining plant water uptake of C. korshinskii, and provide valuable insights for accurately quantifying plant water uptake in the presence of hydrogen isotope offsets between xylem and source water. This study is significant for facilitating the application of the stable hydrogen and oxygen isotope technique worldwide, and for revealing the response mechanism of shrub key ecohydrological and physiological processes to the drought environment in similar climate regions.

The spatio-temporal dynamics of suspended sediment sources based on a novel indexing approach combining Bayesian geochemical fingerprinting with physically-based modelling

Applications of sediment source fingerprinting continue to increase globally as the need for information to support improved management of the sediment problem persists. In our novel research, a Bayesian fingerprinting approach using MixSIAR was used with geochemical signatures, both without and with informative priors based on particle size and slope. The source estimates were compared with a newly proposed Source Sensitivity Index (SSI) and outputs from the INVEST-SDR model. MixSIAR results with informative priors indicated that agricultural and barren lands are the principal sediment sources (contributing ∼5-85% and ∼5-80% respectively during two sampling periods i.e. 2018-2019 and 2021-2022) with forests being less important. The SSI spatial maps (using % clay and slope as informative priors) showed >78% agreement with the spatial map derived using the INVEST-SDR model in terms of sub-catchment prioritization for spatial sediment source contributions. This study demonstrates the benefits of combining geochemical sediment source fingerprinting with SSI indices in larger catchments where the spatial prioritization of soil and water conservation is both challenging but warranted.

[Water Chemical Isotope Characteristics and Water Transformation Relationship in Mongolian Section of the Yellow River Basin]

The Yellow River in Inner Mongolia was selected as the study area in this study. In July (wet season) and October (dry season) of 2021, the acquisition of seasonal rivers, the Yellow River tributaries and precipitation, the Yellow River, Wuliangsuhai, Lake Hasuhai, Lake Daihai, an irrigation canal system, and underground water and sea water samples were collected to test the water chemical composition and hydrogen and oxygen isotopic values of different water types. Using the Piper triplot, Gibbs plot, ion ratio, and MixSIAR model methods, the evolution of water chemistry in the Mongolian section of the Yellow River Basin was analyzed, and the transformation relationship between precipitation, surface water, and groundwater was revealed. The results showed that both groundwater and surface water in the study area were slightly alkaline; the dominant anion in water was Cl-, and the dominant cation was Na+. The main hydrochemical types of surface water were Cl·SO4-Na·Mg and SO4·HCO3-Na·Mg, whereas those of groundwater were Cl·SO4-Na·Mg and SO4·HCO3-Na·Ca. Groundwater Ca2+ and Mg2+ were primarily derived from the dissolution of silicate and evaporite, and surface water Ca2+ and Mg2+ were primarily derived from carbonate karst dissolution and carbonate and sulfuric acid in water participating in the dissolution process of carbonate and sulfide minerals. Na+ and Cl- in different water bodies were all affected by anthropogenic pollution sources. Owing to the seasonal effect, δD and δ18O of surface water and groundwater were higher in the wet season than in the dry season. The results showed that surface water was affected by evaporative fractionation after receiving precipitation recharge, and the groundwater recharge sources were complex. The MixSIAR model revealed that surface water was the main recharge source of groundwater, accounting for 52.4%-62.2% of the total recharge, and atmospheric precipitation was the main recharge source of surface water, accounting for 85.4%-97.1% of the total recharge.

Determining the trophic transfer of metal(loid)s and arsenic speciation in freshwater aquatic organisms by quantifying diet compositions

Bioaccumulation through diet is the predominant source of metal(loid)s in fishes; however, the trophic transfer of metal(loid)s from the diet to aquatic organisms remains largely unclear. In this study, aquatic organisms and five potential food sources (leaf litter, coarse and fine particulate organic matter (CPOM and FPOM, respectively), epilithon and fish) were collected around the Shimen Realgar Mine of China. Stomach content analysis and stable nitrogen and carbon isotope analysis, combined with a new Bayesian mixing model (MixSIAR), were used to quantify diet compositions of aquatic organisms. The δ¹³C and δ¹⁵N values varied among fish sizes and sampling sites and were probably related to the diet shift of aquatic organisms. The MixSIAR modelling results showed that the aquatic organisms' food sources were mainly composed of FPOM (9%-68%) and epilithon (15%-65%), with leaf litter, CPOM and fish accounting for smaller proportions (2%-30%). Concentrations ranged from 0.91 to 1298 mg/kg for As, 0.01-1.30 mg/kg for Cd, 0.12-37.79 mg/kg for Pb, 0.63-1158 mg/kg for Cr, 1.22-411 mg/kg for Cu, 0.82-1772 mg/kg for Mn, 0.31-542 mg/kg for Ni and 21.84-1414 mg/kg for Zn in all the collected samples, including the aquatic organisms and the relevant food sources. The metal(loid) concentrations in the CPOM, FPOM and epilithon were significantly higher than those in aquatic organisms and leaf litter. In addition, the biomagnification factors were all less than 1, indicating a biodilution from diet to freshwater organisms. The predominant As species were organic As in aquatic organisms, while inorganic As was common in their food sources, indicating that As biotransformation occurred within the freshwater food chain.

Identification and contribution of potential sources to atmospheric lead pollution in a typical megacity: Insights from isotope analysis and the Bayesian mixing model

Atmospheric particulate matter (PM) enriched with lead (Pb) has severe irreversible effects on human health. Therefore, identifying the contribution of Pb emission sources is essential for protecting the health of residents. Using the Pb isotopic tracer method, this study explored the seasonal characteristics and primary anthropogenic Pb sources for atmospheric PM in Tianjin in 2019. We calculated the contribution of Pb sources using the end-member and MixSIAR models. The results showed that Pb loaded in PM₁₀ was more abundant in January than in July, and was strongly influenced by meteorological conditions and anthropogenic emissions. The primary Pb sources of the aerosol samples originated from coal combustion and vehicle and steel plant emissions, mainly originating from local Pb emission sources in Tianjin. The PM₁₀-bond Pb in January was influenced by regional transportation and local sources. The MixSIAS model calculated the contribution of coal combustion as approximately 50 %. Compared with that in January, the contribution of coal combustion decreased by 9.6 % in July. Our results indicate that some of the benefits of phased-out leaded gasoline have been short-lived, whereas other industrial activities releasing Pb have increased. Furthermore, the results emphasise the practicability of the Pb isotope tracer source approach for identifying and distinguishing between different anthropogenic Pb inputs. Based on this study, scientific and effective air pollution prevention and control programs can be formulated to provide decision support for the guidance and control of air pollutant emissions.

Effect of rainfall-runoff process on sources and transformations of nitrate using a combined approach of dual isotopes, hydrochemical and Bayesian model in the Dagang River basin

The control of agricultural nitrogen through inflow rivers is crucial for lake aquatic environment conservation, while nitrate is the main form of non-point source pollution of agricultural nitrogen in watersheds. Therefore, understanding the nitrate sources and transformation in the intensive-agricultural region was particularly concerned. Nitrate sources and transformation varied largely during some crucial periods or events. However, low-resolution sampling campaigns increased some uncertainties due to without considering the effect of key driving factors for identifying nitrate transformations and sources. In our study, high-frequency sampling and analysis of nitrogen and oxygen isotope, hydrochemical and Bayesian model was conducted at the Dagang River to capture nitrate sources and transformations and identify its response to rainfall-runoff process at the event scale. In addition, the N cycle process was refined by comparing the variation and relationship of water quality factors and isotopes before, during, and after rainfall. We found that nitrate and major ions derived from similar agricultural activities caused by anthropogenic factors, such as domestic sewage from rural residents and livestock waste, through field survey and principal component analysis. The δ¹⁵N-NO₃^- and δ¹⁸O-NO₃^- in Dagang River ranged from +0.05‰ to +9.94‰ and + 1.49‰ to +11.64‰, respectively. The spatio-temporal variations of nitrate isotopes and hydrochemical compositions of river water suggested that nitrification was the dominant nitrate transformation process although the mixing effect occurred in some periods, especially during, and after the rainfall. The relationship between NO₃^-/Cl^- and Cl^- ratios suggested the occurrence of denitrification in downstream of the river basin after the rainfall. The results of Bayesian model showed that livestock manure and groundwater contributed to the most (66.4 ± 31.9%) nitrate, which indicated the necessity to establish its regulatory policy to avoid the overuse of livestock manure and groundwater in Dagang River. This study benefit for developing concrete and legible management and conservation strategies for decreasing the effect of anthropogenic nitrogen loading on lake eutrophication.

Source tracing of potentially toxic elements in soils around a typical coking plant in an industrial area in northern China

Coking plants are a substantial source of potentially toxic elements (PTEs) in soil. In this study, we examined the concentration of PTEs, the soil physicochemical properties, and the Pb isotopes in the soil inside and around a coking plant in an industrial city in northern China. We analyzed the spatial distribution of PTEs and the pollution risk areas by I_geo index, the enrichment factor (EF), and the Nemerow index, and we quantitatively identified the contribution of PTE pollution sources in the soil on a small- and medium-scale (plant and work section). Our results indicated that the Hg concentration inside the plant and the Cd concentration in the agricultural land around the plant were both relatively high. A comprehensive analysis of the soil in the study area was performed using the positive matrix factorization model and Pb isotope (^206/207Pb, ^208/206Pb) tracing method, based on the MixSIAR model, this analysis indicated that burning coal was the main source of Pb both inside (46.8%) and outside (26.3%) the coking plant. The pollution emission sources with significant influence on the soil outside the coking plant were diesel vehicles (12.5%), gas tanks (12.4%), and coke ovens (11.5%), while the sources inside the plant were quenching sections (11.1%), atmospheric deposition (11.0%), coke oven sections (9.6%), and diesel vehicles (6.1%). The results of PTE pollution risk zoning and Pb isotope tracing indicated that pollution is more serious in the western part of the plant, which is the area where coking and gas production takes place, and the most serious pollution outside the plant is mainly distributed to the southeast. This study provides theoretical and practical data indicating the contribution of industrial enterprises to soil pollution, and will help identify pollution responsibility and the management of pollution sources.

Application of the hydrochemistry, stable isotopes and MixSIAR model to identify nitrate sources and transformations in surface water and groundwater of an intensive agricultural karst wetland in Guilin, China

Karst water as the vital water supply source is generally suffered from NO₃^- contamination in intensive agricultural areas worldwide. Identifying NO₃^- sources and transformations is the key for understanding nitrogen pathways, and also for effectively controlling diffuse NO₃^- pollution. In this study, chemical variables and stable isotopes (δ²H-H₂O, δ¹⁸O-H₂O, δ¹⁵N-NO₃^- and δ¹⁸O-NO₃^-) were measured in 10 surface water (SW) samples and 13 groundwater (GW) samples collected from the Huixian karst wetland, with the application of a Bayesian stable isotope mixing model (MixSIAR) to identified NO₃^- sources and biogeochemical transformations. The results showed that the NO₃^- concentrations ranged from the below detection limit to 117 mg/L, with 30.8% of GW samples obtained from the north central part of the study area exceeding the maximum permissible limit for drinking water, and posing significant non-carcinogenic health risks for native people through drinking water pathway. Moreover, based on characteristics of the hydrochemistry and stable isotopes, different biogeochemical fates were evaluated in SW and GW: nitrification process was a dominant factor in GW, as a result of high NO₃^- levels, and this microbial process was unlikely occurred in SW associated with relatively anaerobic condition and low NO₃^- levels; however, the denitrification might not be a main process of degradation NO₃^- levels throughout the study area. The MixSIAR outputs revealed that the long-term application of synthetic NH₄⁺ fertilizer (36.6%) and soil organic nitrogen (28.0%) were the main contributors to NO₃^- pollution, followed by synthetic NO₃^- fertilizer (16.8%) and domestic sewage and manure (15.1%), whereas NO₃^- in precipitation (3.44%) played a less important role. Additionally, NO₃^- concentration was significantly influenced by agricultural activities rather than NO₃^- source's contribution between SW and GW. This work suggests that synthetic NH₄⁺ fertilizer should be the primary target for control to prevent further NO₃^- pollution of the karst groundwater.

Nitrate migration and transformations in groundwater quantified by dual nitrate isotopes and hydrochemistry in a karst World Heritage site

Karst aquifers are extremely vulnerable to pollution, including from nitrate. This research advances a systematic evaluation of water quality dynamics and processes in a relatively pristine karst flow system impacted by seasonal tourism, in particular migration and transformation of nitrate. Water samples from the Shuifang Spring basin (Jinfoshan Karst World Heritage Site, Chongqing, China) were collected for analysis of ion concentrations and dual nitrate isotopes. The principal sampling sites included untreated tap water from the Jinfoshan Holiday Hotel (JHH), JHH septic system effluent, discharged effluent at Sinkhole #1, and groundwater at Shuifang Spring (SFS, the resurgence of the basin). Sampling from April 2017 through April 2018 showed that higher nitrate concentrations were observed at SFS during busy tourist seasons that occurred in the summer drought and winter dry seasons. Between the septic system and SFS, nitrate was diluted by rainwater and infiltration from the matrix regions of the karst aquifer in the wet season. The MixSIAR model quantitatively revealed that nitrate from manure and sewage (M&S), soil organic nitrogen (SON) and atmospheric precipitation (AP) contributed roughly 55%, 41% and 4% on average, respectively. The mixture of effluent and oxygen-rich tap water from the JHH underwent nitrification at the septic system and Sinkhole #1. After the nitrification and decomposition of organic matter with oxygen consumption in the effluent, denitrification was the dominant response to nitrate attenuation in the effluent of Sinkhole #1 during peak tourism periods. The septic effluent sinking into the aquifer deteriorated the groundwater quality, which was controlled by nitrification in the summer drought and winter dry seasons. In contrast, microorganisms in groundwater assimilated nitrate from effluent in the wet season. Because of the relatively pristine ambient background conditions and seasonal tourism, these results can inform interpretation of background and impacted nitrogen processes in a range of karst settings.

Identifying the composition of commercial Brazilian cat food with stable isotopes of carbon and nitrogen

Considering the increasing pet owner's concern about the food their pets are consuming, in this study we investigated the origin of the main ingredients in wet and dry foods produced in Brazil using stable isotope ratios of carbon and nitrogen. We concluded that chicken and pork seem to be the dominant ingredients in most of the samples, with larger proportions in wet cat food. Even in pet foods showing 'beef' as the main ingredient on the label, we found a low proportion of bovine products in both wet and dry cat foods. Comparing the contribution of plant-derived products (C₃ and C₄ plants) and animal-derived products (chicken-pork, bovine and fish), approximately 21 % of cat foods had more than 30 % of ingredients with plant origin in their composition. The high amount of plant-derived products in cat foods found here raises the question whether this should be mentioned on package labels.

Identifying water sources used by alpine riparian plants in a restoration zone on the Qinghai-Tibet Plateau: Evidence from stable isotopes

Riparian zones are vulnerable to water regimes, which alter soil water availability and impact the persistence of riparian plants. However, little is known about the water use patterns of alpine riparian shrubs (e.g., Myricaria squamosa) in response to changes in soil water availability on the Qinghai-Tibet Plateau. This study examined the water-use patterns of M. squamosa along a zone of gradual degradation (light, moderate, and severe), located in the lower reaches of the Shaliu River in the Qinghai Lake watershed. Stable water isotopes (δ²H and δ¹⁸O) in xylem water, soil water and groundwater, as well as leaf water potential were monitored during the growing seasons from 2012 to 2013, and quantified the water-use proportions via MixSIAR model. Results showed significant differences in the isotopic signatures of M. squamosa from the light, moderate, and severe degradation sites across seasons, suggesting that M. squamosa exploited different water sources. MixSIAR results also revealed that M. squamosa used high proportions of shallow soil water in the light degradation site (35.4%) compared with the severe degradation sites (13.4%). By contrast, M. squamosa exhibited an ability to shift its water sources and to rely more on groundwater in the severe degradation site across seasons. The contrasting water-use patterns of M. squamosa along the gradual degradation zone were closely linked with the distributions of active root zones when competing for water. Higher predawn leaf water potential (Ψ_pd) of M. squamosa (mean Ψ_pd = -2.29 ± 0.7 MPa) was found in the light degradation site and lower Ψ_pd values in the severe degradation site (mean Ψ_pd = -3.3 ± 0.8 MPa), suggesting that M. squamosa depended on a high degree of flexible plasticity in water use to alleviate water stress along the gradual degradation. These results linked to water-use patterns and ecophysiological characteristics (e.g., Ψ_pd) of plants responding to changes in available water are important for informing decision-making management strategies designed to prevent ecological degradation.

Using stable isotopes to determine seasonal variations in water uptake of summer maize under different fertilization treatments

Fertilization and water both affect root water uptake in the nutrient and water cycle of the Soil-Plant-Atmosphere-Continuum (SPAC). In this study, dual stable isotopes (D and (18)O) were used to determine seasonal variations in water uptake patterns of summer maize under different fertilization treatments in Beijing, China during 2013-2014. The contributions of soil water at different depths to water uptake were quantified by the MixSIAR Bayesian mixing model. Water uptake was mainly sourced from soil water in the 0-20cm depth at the seeding (67.7%), jointing (60.5%), tasseling (47.5%), dough (41.4%), and harvest (43.9%) stages, and the 20-50cm depth at the milk stage (32.8%). Different levels of fertilization application led to considerable differences in the proportional contribution of soil water at 0-20cm (6.0-58.5%) and 20-50cm (6.1-26.3%). There was little difference of contributions in the deep layers (50-200cm) among treatments in 2013, whereas differences were observed in 50-90cm at the milk stage and 50-200cm at the dough stage during 2014. The main water uptake depth was concentrated in the upper soil layers (0-50cm) during the wet season (2013), whereas a seasonal drought in 2014 promoted the contribution of soil water in deep layers. The contribution of soil water was significantly and positively correlated with the proportions of root length (r=0.753, p<0.01). The changes of soil water distribution were consistent with the seasonal variation in water uptake patterns. The present study identified water sources for summer maize under varying fertilization treatments and provided scientific implications for fertilization and irrigation management.