Regional Calibration With Isotope Tracers Using a Spatially Distributed Model: A Comparison of Methods

Abstract

Accurate representation of flow sources in process-based hydrologic models remains challenging for remote, data-scarce regions. This study applies stable isotope tracers (18 O and 2 H) in water as auxiliary data for the calibration of the isoWATFLOOD ™ model. The most efficient method of those evaluated for introducing isotope data into model calibration was the PA-DDS multiobjective search algorithm. The compromise solutions incorporating isotope data performed slightly inferior in terms of streamflow simulation compared to the calibrated solution using streamflow data only. However, the former solution outperformed the latter one in terms of isotope simulation. Approximation of the model parameter uncertainty into internal flow path partitioning was explored. Inclusion of isotope error facilitated a broader examination of the total parameter space, resulting in significant differences in internal storage and flow paths, most significantly for soil storage and evapotranspiration loss. Isotope-optimized calibration reduced evaporation rates and increased soil moisture content within the model, impacting soil water velocity but not streamflow celerity. Flow-only calibration resulted in artificially narrow model prediction bounds, significantly underestimating the propagation of parameter uncertainty, while isotope-informed calibrations yielded more reliable and robust bound on model predictions. Our findings demonstrate that the accuracy of a complex, spatially distributed, and process-based model cannot be judged from one summative flow-based model performance evaluation metric alone. Plain Language Summary To predict how water moves through watersheds and adequately anticipate hydrologic response to a changing climate, we depend on numerical models to estimate transport and storage of water on and below Earth's surface. These models are only as good as the observation data we have to test them against, which can be insufficient for many remote parts of the world. Our study introduces a new source of data (i.e., isotope tracers, a fingerprint of water's origin) into a model to improve parameter estimation and overall reliability. The largest benefit of adding these data is that it nudges the model into a more realistic distribution of water storage and flow paths to the rivers, particularly soil storage and evaporation loss. This is important for improving the reliability of model projections under changing climates and to better understand how the distribution of water on Earth is changing.

Catchment Functioning Under Prolonged Drought Stress: Tracer‐Aided Ecohydrological Modeling in an Intensively Managed Agricultural Catchment

“…Moreover, other model performances at catchment (e.g., discharge) and plot (e.g., soil moisture) scales were maintained (and even slightly better for surface temperature, Table 2). This supports the argument that the added tracer information facilitated a more robust capturing of internal processes (e.g., storage-flux interactions, Figure 4; flow pathways, Figure 5), which is consistent with other studies that evaluated the value of isotope data for model calibration (He et al, 2019;Holmes et al, 2020;.…”

“…Large variations in information content, as well as associated uncertainty, reside in different data types and different periods of data collection (Beven & Smith, 2015; Westerberg & McMillan, 2015). Therefore, multi‐criteria calibration based on contrasting data sources at different scales is important, especially for spatially distributed models (Fatichi et al., 2016; Holmes et al., 2020; Kuppel et al., 2018a). In particular, information on stable isotopes of water (deuterium and oxygen‐18 ratios; and , respectively) has been increasingly integrated in ecohydrological modeling, from which storage‐flux interactions, precipitation partitioning and surface water dynamics can be tracked and verified (Birkel & Soulsby, 2015; McGuire & McDonnell, 2015; Tetzlaff et al., 2015).…”

Quantifying the effects of land use and model scale on water partitioning and water ages using tracer-aided ecohydrological models

“…The feedback between ecology and hydrology is, however, strongly scale dependent, with controls on interactions vastly different across space and time (Fatichi et al, 2015). The interdependency of models on temporal and spatial scales often confounds identifiability of hydrological processes due to emergent behaviour, nonlinearity of parameter interactions, and aggregation effects at coarser resolutions when models are applied at larger scales (Wood et al, 1988;Blöschl and Sivapalan, 1995;Horritt and Bates, 2001;Samaniego et al, 2017). Many of the advancements in addressing difficulties in model scaling have focused on discharge (Samaniego et al, 2017) or soil moisture (Vereecken et al, 2008) due to limited alternative data and their information as proxies for large-scale water availability and atmospheric exchange.…”

Estimating Reservoir Inflow and Outflow From Water Level Observations Using Expert Knowledge: Dealing With an Ill‐Posed Water Balance Equation in Reservoir Management

“…Previous research has focused on investigating how many observations or how much variability from peak to base flows is required to obtain acceptable calibration accuracy rather than gleaning additional data and information from other partially known variables. Recent studies have attempted to derive information from various sources of data, such as remotely sensed products (e.g., Dembélé et al., 2020; Huang et al., 2020; Nijzink et al., 2018), isotope tracers (e.g., Holmes et al., 2020), and crowdsourced data (e.g., Avellaneda et al., 2020; Weeser et al., 2019), and they have found that the derived information can improve the prediction accuracy of hydrological processes. Reservoir water levels are direct observations, but they have not often been used for reservoir analysis, presumably due to the lack of a framework for the efficient use of such data.…”