Revising Common Approaches for Calibration: Insights From a 1‐D Tracer‐Aided Hydrological Model With High‐Dimensional Parameters and Objectives

Water Resources Research, Journal Year: 2025, Volume and Issue: 61(1)

Published: Jan. 1, 2025

Abstract The dimensionality of parameters and objectives has been increasing due to the accelerating development models monitoring network, which brings potential challenges for calibration. In this study, two common philosophies multi‐objective optimisation in hydrology (the use aggregated scalar criterion or vector functions) were revisited with different sampling strategies: (a) random sampling, (b) DiffeRential Evolution Adaptive Metropolis (DREAM as an example function), (c) Non‐Dominated Sorting Genetic Algorithm II (NSGA‐II Pareto‐based optimisation). By testing ability algorithms simultaneously capture soil moisture water isotopes at three depths under four vegetation covers, we found performed poorly matching observations its inability explore high‐dimensional parameter space. DREAM, contrast, could provide efficient convergence informal likelihood functions, but choice formal function is difficult lack knowledge about model residuals, leading poor performance. NSGA‐II effective after aggregating ≤4, failed when calibrating against all 24 objectives. Overall, both approaches are challenged by dimensionality, it generally requires a degree trial‐and‐error before achieving successful This suggests more flexible way describe residuals (e.g., defining limits acceptability). Alternatively, improvements be made using ensemble represent system (instead “best” model) given average calibrated usually better than any individual model.

Language: Английский

---

Seasonal and Inter‐Annual Dynamics in Water Quality and Stream Metabolism in a Beaver‐Impacted Drought‐Sensitive Lowland Catchment

Hydrological Processes, Journal Year: 2025, Volume and Issue: 39(2)

Published: Feb. 1, 2025

ABSTRACT Increasing drought frequency and severity from climate change are causing streamflow to become increasingly intermittent in many areas. This has implications for the spatio‐temporal characteristics of water quality regimes which need be understood terms risks provision clean public supplies instream habitats. Recent advances sensor technology allow reliable accurate high‐resolution monitoring a growing number parameters. Here, we continuously monitored suite parameters over 3 years an stream network eutrophic, lowland Demnitzer Millcreek catchment, Germany. We focused on effects wetland systems impacted by beaver dams diurnal, seasonal inter‐annual variation dynamics at two sites, upstream downstream these wetlands. then used data model metabolism. Dissolved oxygen pH were higher wetlands, while conductivity, turbidity, chlorophyll phosphorous concentrations downstream. found clear diurnal cycling dissolved both sites. These correlated with hydroclimatic changes metabolism, becoming pronounced as temperatures increased flows decreased spring summer. Upstream wetlands this corresponded rapidly heterotrophic modelled Gross Primary Production (GPP) was exceeded Ecosystem Respiration (ER). Downstream, where GPP lower, usually strongly prone hypoxic conditions (i.e., insufficient oxygen) before ceased coincided lower velocities deeper channels Seasonal variations mainly correlate factors (particularly temperature) their influence streamflow. study highlights that heterotrophy hypoxia rivers central Europe is important feature streams agricultural landscapes continue leaching nutrients. insights contribute evidence base understanding how will affect quantity rural resources presence beavers requires management responses.

Language: Английский

---

DREAM_(LoAX): Simultaneous Calibration and Diagnosis for Tracer‐Aided Ecohydrological Models Under the Equifinality Thesis

Water Resources Research, Journal Year: 2025, Volume and Issue: 61(3)

Published: March 1, 2025

Abstract The Limits of Acceptability approach has been demonstrated to be an effective conditioning tool due its capacity consider epistemic uncertainty. However, application faces two challenges—the low efficiency when random sampling is used and the difficulty in setting limits prior calibration. Here algorithm DREAM (LoAX) was developed added GLUE framework. As extension (LoA) Vrugt Beven (2018), https://doi.org/10.1016/j.jhydrol.2018.02.026 , it evaluates model performance based on limit boundaries, thus inherits merits framework (explicit consideration errors). Moreover, importance initial choice strongly reduced by allowing iterative evolution historical performance. By testing a series examples (including high‐dimensional numeric example, single‐objective hydrological multi‐objective example) with or without error‐free assumption using synthetic real observations, search locate acceptable models demonstrated. also shows comparable . More importantly, provides real‐time diagnostic information regarding (at which timestep) where (for objective) how (to direction extent) fails uncertainty pronounced, potential sources data flaws structure identified. In this context, not only useful tool, but learning for development improved modeling.

Language: Английский

---

Recent Developments and Emerging Challenges in Tracer‐Aided Modeling

Wiley Interdisciplinary Reviews Water, Journal Year: 2025, Volume and Issue: 12(2)

Published: March 1, 2025

ABSTRACT During the last decade, tracer‐aided hydrological models (TAMs) have been applied in numerous studies and successfully evolved for different purposes. Such confirmed value of tracer data modeling, offering insights into internal storages, water sources, flow pathways, mixing processes, ages, which cannot be derived from hydrometric alone. The direct coupling tracers flux tracking balance can reduce model uncertainty through increased biogeochemical process knowledge. More specifically, such simultaneously capture celerity responses with velocities (and age) particles. As a result availability high‐resolution characterizing functioning across Critical Zone entire landscapes, together rapid improvement computing capacity, four major advances reshaped capability TAMs, we review this paper: (1) enhanced representation spatial heterogeneity, (2) more explicit conceptualization ecohydrological partitioning, (3) application to larger catchment scales, (4) incorporation non‐conservative coupled quality modeling. However, persistent challenges also emerged, particularly relation acquisition, mismatches between information content scale application, uncertainties structures, as well adaptation multi‐criteria calibration. In review, recent remaining TAMs summarized discussed particular focus on conservative models.

Language: Английский

---

The Unexploited Treasures of Hydrological Observations Beyond Streamflow for Catchment Modeling

Wiley Interdisciplinary Reviews Water, Journal Year: 2025, Volume and Issue: 12(2)

Published: March 1, 2025

ABSTRACT While measured streamflow is commonly used for hydrological model evaluation and calibration, an increasing amount of data on additional variables available. These have the potential to improve process consistency in modeling consequently predictions under change, as well data‐scarce or ungauged regions. Here, we show how these beyond are currently calibration. We consider storage flux variables, namely snow, soil moisture, groundwater level, terrestrial water storage, evapotranspiration, altimetric level. aim at summarizing state‐of‐the‐art providing guidance use Based a review current literature, summarize observation methods uncertainties available sets, challenges regarding their implementation, benefits consistency. The focus catchment studies with study areas ranging from few km 2 ~500,000 . discuss implementing alternative that related differences spatio‐temporal resolution observations models, variable‐specific features, example, discrepancy between observed simulated variables. further advancements required deal integrate multiple, potentially inconsistent datasets. increased improvement shown by most reviewed often come cost slight decrease performance.

Language: Английский

---

Increasing parameter identifiability through clustered time-varying sensitivity analysis

Environmental Modelling & Software, Journal Year: 2024, Volume and Issue: 181, P. 106189 - 106189

Published: Aug. 19, 2024

Language: Английский

---

Can temperature be a low-cost tracer for modelling water age distributions in a karst catchment?

Journal of Hydrology, Journal Year: 2024, Volume and Issue: unknown, P. 131947 - 131947

Published: Sept. 1, 2024

Language: Английский

---

Hydrological connectivity drives intra- and inter-annual variation in water quality in an intermittent stream network in a mixed land use catchment under drought

Journal of Hydrology, Journal Year: 2024, Volume and Issue: 648, P. 132420 - 132420

Published: Nov. 26, 2024

Language: Английский

---