Data assimilation (DA) is used to integrate artificial intelligence including machine learning (AI-ML) and process-based models to produce a dynamic operational water balance tool for groundwater management. The management tool is a three-step calculation. In the first step, a traditional process-based water budget model provides forward model predictions of aquifer storage from meteorological observations, estimates of pumping and diversion discharge, and estimates of recharge. A Kalman filter-based DA approach is the second step and generates updated storage volumes by combining a trained AI-ML model, providing replacement 'measurements' for missing observations, with forward model predictions. The third 'correction' step uses modified recharge and pumping, adjusted to account for the difference between Kalman update storage and forward model predicted storage, in forward model re-simulation to approximate updated storage volume. Use of modified inputs in the correction provides a mass conservative water budget framework based on AI-ML predictions. Pumping and recharge values are uncertain and unobserved in the study region and can be adjusted without contradicting measurements.

Data assimilation (DA) makes the best combination of model simulation results and observed, or measured, values. Ensemble methods are a form of DA that generates multiple equally good, or equally calibrated, models using a description of model and observation uncertainty. Uncertainty is lack of knowledge. The collection of equally good, in the presence of uncertainty, models is an ensemble of models. An observation error model provides the means to describe the amount of uncertainty in model simulation results and in observed values as part of assimilation. Model-related uncertainty comes from model representation limitations created by differences between what the model represents, or simulates, and by what is measured to make an observation. Observation uncertainty comes from observation error. When an observed value is calculated or estimated, rather than measured, additional uncertainty is generated by the estimation procedure.

An observation error model is developed and presented for river discharge observations made at a stream gauging station using a measured water depth value with a derived rating curve to calculate discharge from observed water depth. A rating curve is a poor hydrodynamics model. Consequently, large estimation errors are expected for river discharge calculated using a rating curve, which generates correspondingly large amounts of observation uncertainty for assimilation. Uncertainty is propagated through DA to the spread, or variability, of model outcomes provided by the ensemble of models. When assimilating simulation results and data with significant uncertainty, the goal of assimilation is to optimize the bias-variance tradeoff and thus the spread of ensemble outcomes. Optimizing this tradeoff involves limiting the amount of uncertainty as much as possible to make informed decisions while including sufficient uncertainty to avoid overfitting. The risk from overfitting is production of biased model outcomes and spurious decision support.

Artificial intelligence (AI) is the effort to automate intellectual tasks normally performed by humans, and it includes machine learning (ML) and deep learning (DL) approaches. AI-ML is based on statistical learning, which tries to learn statistics-based rules for data analyses from known examples of inputs and corresponding outcomes. Data sets that are noisy, include significant uncertainty, and have extreme values hinder statistical learning. ML and DL aquifer recharge predictors are developed to: (1) examine prediction skill when trained using noisy and uncertain data and (2) identify advantages of AI-ML relative to traditional physics- and process-based calculations. Recharge was selected as the learning outcome because it is not observed and is inherently uncertain. A common-sense baseline is developed and implemented to account for uncertainty and noise in AI-ML predictions. The baseline provides a lower goodness-of-fit threshold that identifies when trained AI-ML generates prediction skill and an upper goodness-of-fit threshold above which the AI-ML is learning to reproduce noise and bias in the training data set (and is likely overfitting). Identified advantages for AI-ML (relative to physics- or process-based calculations) are the ability to use dimensionless trends for features and to represent a complex scenario with the same level of effort as for a simple case.

Weather attribution estimates the current and near future likelihood for a recently observed extreme weather event, like a drought or hurricane. It uses climate models, weather prediction models, and observed weather to determine how much more likely the observed event is today relative to the recent past, like the 1990s and 2000s. In this study, a statistical weather generator (WG) creates synthetic sequences of future precipitation, temperature, and potential evapotranspiration that represent the increased likelihood for three-month severe drought. An independent weather attribution study identified that three-month severe drought is five times more likely to occur today relative to recent historical conditions. The WG-simulated conditions portray a near future where historical extreme and severe drought are significantly more likely to occur. The climate description produced by this WG is representative of the weather attribution study and is significantly hotter, with lower expected soil moisture than the future climate description obtained from global circulation, i.e., climate, model (GCM) simulation results (by themselves).

A methodology is presented for predicting impacts and risks to water resources, at the watershed scale, from somewhat unknown future climate. It is then applied to estimate impacts to a semi-arid watershed in Texas. Because all models of water movement and storage in watersheds provide estimates (and best guesses), rather than absolute answers, and because the specifics of future weather are unknown, this approach uses likelihoods (or probabilities) for relative change in magnitude, ∆, between future and historical precipitation, evapotranspiration, storm runoff, and aquifer recharge to evaluate future risk to water availability. Projected (future) climate trends for the study site from climate models are a 3 ˚C increase in average temperature, which means that the potential for evapotranspiration will increase, no significant change in average annual precipitation, which means that there generally will not be more water available for evaporation, and a semi-arid classification from 2011–2100. Future precipitation is projected as unchanged for typical conditions. Consequently, no significant change is estimated for evapotranspiration, runoff, or recharge for average conditions. With expectations for significant temperature increase, an increase in the amount of rainfall is needed to increase evapotranspiration, runoff, and recharge. Increases in rainfall during infrequent large storms are included in the analysis for future conditions, which produces increased water availability during infrequent extreme events but does not change expectations for average conditions.

Climate change and changes to land use and land cover (LULC) both impact water resources, and they have interacting influences on the amount of water available for management and consumption. The framework for the assessment of relative risk to watershed-scale water resources from systemic changes presented in 'Projecting Climate Change Impacts to Watershed Water Resources' is used again to predict combined climate and LULC change impacts from 2011–2100 for the same semi-arid watershed in Texas. In the application, an increase in impervious area from economic development is the LULC change. It generates a 1.1 times increase in average water availability, relative to future climate trends, from increased runoff and decreased evapotranspiration.