Groundwater Simulations and Uncertainty Analysis Using MODFLOW and Geostatistical Approach with Conditioning Multi-Aquifer Spatial Covariance

Stochastic nitrate simulation under hydraulic conductivity uncertainty of an agricultural basin aquifer at Eastern Thessaly, Greece

Lake Karla Watershed is an agricultural basin characterized by intense agricultural activities, which lead to quantitative and qualitative aquifer degradation. The exploitation of non-renewable groundwater resources and nitrate contamination are the major threats to aquifer sustainability. Groundwater resources cover mainly irrigation and domestic water needs. Hence, the simulation of groundwater resources is necessary to (a) determine the quantity available for water supply and (b) estimate probable nitrate contamination to propose subsequent remediation techniques to protect public health. Furthermore, the aquifer's heterogeneity as well as the lack of hydraulic conductivity data, in a large-scale study area, creates uncertainty regarding groundwater flow and nitrate transport simulation. Deterministic modelling approaches for spatially distributed nitrate concentration simulation could not estimate the contamination risk, since it can address only one realization of the aquifer. This study estimates the effect of hydraulic conductivity uncertainty on the simulation of groundwater nitrate concentration. The proposed framework uses Geostatistical Sequential Gaussian Simulation (SGSIM) for the generation of equally probable realizations of the spatial distribution of hydraulic conductivity. It includes the application of a modelling system based on the following inter-linked models: a rainfall-runoff model, a reservoir operation model, a lake-aquifer interaction model, a groundwater flow model, and a nitrate transport and dispersion model. The last two models simulate the multiple realizations of aquifer's groundwater flow and nitrate concentration. Furthermore, two analyses (a statistical and a threshold analysis) are employed to estimate the exceedance probability of the nitrate concentrations and their spatial extent. The results indicate that hydraulic conductivity uncertainty does affect the simulation of nitrate concentration and that nitrate concentrations will most probably exceed the thresholds in areas where groundwater is extracted for domestic use.

Assessing thermodynamic parameter sensitivity for simulating temperature responses of soil nitrification

Soil nitrification responses to temperature have major implications for the global nitrogen cycle. Temperature sensitivity of soil nitrification has been modeled using several mathematical models, yet the extent to which model-generated thermodynamic parameters are accurate and sensitive in describing temperature sensitivity is unclear. In this study, we performed global sensitivity analysis to identify the key thermodynamic parameters that are most influential when simulating the temperature response of the soil nitrification potential (NP) across two different temperature gradients (4-40 °C and 20-45 °C) which are imposed upon sixteen different soils with square root growth (SQRT) and macromolecular rate theory (MMRT) models. We found that two thermodynamic parameters stand out as moderately to highly sensitive, and are uniquely identifiable in each model, regardless of the temperature range. The minimum and maximum measured temperatures seem to have no impact on the list of sensitive parameters but do influence the parameter ranges, especially for the SQRT model. However, parameters that control the minimum temperature and curvature of the NP response curve (Tmin and ΔC‡P) were found to have little to no sensitivity to SQRT and MMRT model outputs, respectively. We show that the parameter sensitivity and range of measured temperatures influence the complementary model's ability to describe the temperature sensitivity of soil nitrification. Our proposed framework enhances the accurate interpretation of existing thermodynamic parameters that explain the temperature sensitivity of soil biochemical processes, and provides methodological recommendations for future temperature sensitivity studies.