An official website of the United States government
Abstract. Deep learning (DL) rainfall–runoff models outperform conceptual, process-based models in a range of applications. However, it remains unclear whether DL models can produce physically plausible projections of streamflow under climate change. We investigate this question through a sensitivity analysis of modeled responses to increases in temperature and potential evapotranspiration (PET), with other meteorological variables left unchanged. Previous research has shown that temperature-based PET methods overestimate evaporative water loss under warming compared with energy budget-based PET methods. We therefore assume that reliable streamflow responses to warming should exhibit less evaporative water loss when forced with smaller, energy-budget-based PET compared with temperature-based PET. We conduct this assessment using three conceptual, process-based rainfall–runoff models and three DL models, trained and tested across 212 watersheds in the Great Lakes basin. The DL models include a Long Short-Term Memory network (LSTM), a mass-conserving LSTM (MC-LSTM), and a novel variant of the MC-LSTM that also respects the relationship between PET and evaporative water loss (MC-LSTM-PET). After validating models against historical streamflow and actual evapotranspiration, we force all models with scenarios of warming, historical precipitation, and both temperature-based (Hamon) and energy-budget-based (Priestley–Taylor) PET, and compare their responses in long-term mean daily flow, low flows, high flows, and seasonal streamflow timing. We also explore similar responses using a national LSTM fit to 531 watersheds across the United States to assess how the inclusion of a larger and more diverse set of basins influences signals of hydrological response under warming. The main results of this study are as follows: The three Great Lakes DL models substantially outperform all process-based models in streamflow estimation. The MC-LSTM-PET also matches the best process-based models and outperforms the MC-LSTM in estimating actual evapotranspiration. All process-based models show a downward shift in long-term mean daily flows under warming, but median shifts are considerably larger under temperature-based PET (−17 % to −25 %) than energy-budget-based PET (−6 % to −9 %). The MC-LSTM-PET model exhibits similar differences in water loss across the different PET forcings. Conversely, the LSTM exhibits unrealistically large water losses under warming using Priestley–Taylor PET (−20 %), while the MC-LSTM is relatively insensitive to the PET method. DL models exhibit smaller changes in high flows and seasonal timing of flows as compared with the process-based models, while DL estimates of low flows are within the range estimated by the process-based models. Like the Great Lakes LSTM, the national LSTM also shows unrealistically large water losses under warming (−25 %), but it is more stable when many inputs are changed under warming and better aligns with process-based model responses for seasonal timing of flows. Ultimately, the results of this sensitivity analysis suggest that physical considerations regarding model architecture and input variables may be necessary to promote the physical realism of deep-learning-based hydrological projections under climate change.
more »
« less
Modeling tree radial growth in a warming climate: where, when, and how much do potential evapotranspiration models matter?
Abstract Process-based models of tree-ring width are used both for reconstructing past climates and for projecting changes in growth due to climate change. Since soil moisture observations are unavailable at appropriate spatial and temporal scales, these models generally rely on simple water budgets driven in part by temperature-based potential evapotranspiration (PET) estimates, but the choice of PET model could have large effects on simulated soil moisture, moisture stress, and radial growth. Here, I use four different PET models to drive the VS-Lite model and evaluate the extent to which they differ in both their ability to replicate observed growth variability and their simulated responses to projected 21st century warming. Across more than 1200 tree-ring width chronologies in the conterminous United States, there were no significant differences among the four PET models in their ability to replicate observed radial growth, but the models differed in their responses to 21st century warming. The temperature-driven empirical PET models (Thornthwaite and Hargreaves) simulated much larger warming-induced increases in PET and decreases in soil moisture than the more physically realistic PET models (Priestley–Taylor and Penman–Monteith). In cooler and more mesic regions with relatively minimal moisture constraints to growth, the models simulated similarly small reductions in growth with increased warming. However, in dry regions, the Thornthwaite- and Hargreaves-driven VS-Lite models simulated an increase in moisture stress roughly double that of the Priestley–Taylor and Penman–Monteith models, which also translated to larger simulated declines in radial growth under warming. While the lack of difference in the models’ ability to replicate observed radial growth variability is an encouraging sign for some applications (e.g. attributing changes in growth to specific climatic drivers), the large differences in model responses to warming suggest that caution is needed when applying the temperature-driven PET models to climatic conditions with large trends in temperature.
more »
« less
- Award ID(s):
- 2001753
- PAR ID:
- 10332656
- Date Published:
- Journal Name:
- Environmental Research Letters
- Volume:
- 16
- Issue:
- 8
- ISSN:
- 1748-9326
- Page Range / eLocation ID:
- 084017
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
Abstract Drought is projected to become more severe and widespread as global warming continues in the 21 st century, but hydroclimatic changes and their drivers are not well examined in the latest projections from the Phase Six of the Coupled Model Inetercomparison Project (CMIP6). Here, precipitation (P), evapotranspiration (E), soil moisture (SM), and runoff (R) from 25 CMIP6 models, together with self-calibrated Palmer Drought Severity Index with Penman-Monteith potential evapotranspiration (scPDSIpm), are analyzed to quantify hydroclimatic and drought changes in the 21 st century and the underlying causes. Results confirm consistent drying in these hydroclimatic metrics across most of the Americas (including the Amazon), Europe and the Mediterranean region, southern Africa, and Australia; although the drying magnitude differs, with the drying being more severe and widespread in surface SM than in total SM. Global drought frequency based on surface SM and scPDSIpm increases by ~25%–100% (50%–200%) under the SSP2-4.5 (SSP5-8.5) scenario in the 21 st century together with large increases in drought duration and areas, which result from a decrease in the mean and flattening of the probability distribution functions of SM and scPDSIpm; while the R-based drought changes are relatively small. Changes in both P and E contribute to the SM change, whereas scPDSIpm decreases result from ubiquitous PET increases and P decreases over subtropical areas. The R changes are determined primarily by P changes, while the PET change explains most of the E increase. Inter-model spreads in surface SM and R changes are large, leading to large uncertainties in the drought projections.more » « less
-
Abstract Drylands are highly vulnerable to climate change due to their fragile ecosystems and limited ability to adapt. In contrast to the global drying after tropical volcanic eruptions shown previously, we demonstrate that large tropical volcanic eruptions can induce significant two-year hydroclimatic wetting over drylands by employing the last millennium simulations. During this wetting period, which extends from the first to the third boreal winter after the eruption, several hydroclimatic indicators, such as self-calibrating Palmer Drought Severity Index based on the Penman-Monteith equation for potential evapotranspiration (scPDSIpm), standard precipitation evapotranspiration index (SPEI), aridity index (AI), top-10cm soil moisture (SM10cm), and leaf area index (LAI), show significant positive anomalies over most drylands. The primary contribution to the wetting response is the potential evapotranspiration (PET) reduction resulting from dryland surface cooling and reduced solar radiation, as well as a weak contribution from increased precipitation. The latter is due to the wind convergence into drylands caused by slower tropical cooling compared to drylands. The wetting response of drylands to volcanic eruptions also demonstrates some benefits over the global hydrological slowdown resulting from stratospheric aerosol injection, which replicates the cooling effects of volcanic eruptions to address global warming.more » « less
-
Severe and persistent 21st-century drought in southwestern North America (SWNA) motivates comparisons to medieval megadroughts and questions about the role of anthropogenic climate change. We use hydrological modeling and new 1200-year tree-ring reconstructions of summer soil moisture to demonstrate that the 2000–2018 SWNA drought was the second driest 19-year period since 800 CE, exceeded only by a late-1500s megadrought. The megadrought-like trajectory of 2000–2018 soil moisture was driven by natural variability superimposed on drying due to anthropogenic warming. Anthropogenic trends in temperature, relative humidity, and precipitation estimated from 31 climate models account for 47% (model interquartiles of 35 to 105%) of the 2000–2018 drought severity, pushing an otherwise moderate drought onto a trajectory comparable to the worst SWNA megadroughts since 800 CE.more » « less
-
This study employs a pseudo–global warming approach to investigate precipitation changes from a mesoscale convective system (MCS) on 14 May 2018 over the eastern United States. An Appalachian-Mountain-crossing MCS is simulated for historical, mid-twenty-first century (2045–54), and late-twenty-first century (2090–99) climate scenarios. For experiments using ensemble-mean perturbations in atmospheric, soil, and oceanic variables derived from 34 general circulation models, MCS precipitation diminishes by 25%in the midcentury and 65%in the late century. Experiments testing the sensitivity to these variables separately reveal that atmospheric variables primarily drive precipitation changes. Additional sensitivity experiments quantify MCS responses to temperature, moisture, and wind perturbations separately, with the magnitude of perturbations stratified as low, moderate, or high. Experiments highlight the dominant though contrasting roles of the thermodynamic variables. In midcentury, temperature increases lead to reductions in rainfall rates by up to 74.3%, while increased moisture raises rainfall rates by 75.1%. In the late century, the MCS fails to initiate for temperature perturbations of all magnitudes. Rainfall rate and precipitation area substantially increase with larger moisture perturbations, while the frequency of heavy (95th percentile) and extreme (99th percentile) precipitation increases more than 100%, with minimal changes in precipitation rate. Finally, ensemble-mean perturbations are added to all variables, except for temperature or moisture, to which either a low or high perturbation is added. MCSs are robust when low-temperature or high-moisture perturbations are included, though they fail to initiate for low-moisture and high-temperature perturbations, highlighting the challenges in projecting future MCS behavior.more » « less
- Free Publicly Accessible Full Text
-
Accepted Manuscript1.0
- Journal Article:
- https://doi.org/10.1088/1748-9326/ac1292
