Application of Postprocessing to Watershed-Scale Subseasonal Climate Forecasts over the Contiguous United States

Abstract

Subseasonal to seasonal (S2S) climate forecasting has become a central component of climate services aimed at improving water management. In some cases, operational S2S climate predictions are translated into inputs for follow-on analyses or models, whereas the S2S predictions on their own may provide for qualitative situational awareness. At the spatial scales of water management, however, S2S climate forecasts often suffer from systematic biases, and low skill and reliability. This study assesses the potential to improve S2S forecast skill and salience for watershed applications through the use of postprocessing to harness skills in large-scale fields from the global climate model forecast outputs. To this end, the components-based technique—partial least squares regression (PLSR)—is used to improve the skill of biweekly temperature and precipitation forecasts from the Climate Forecast System version 2 (CFSv2). The PLSR method forms predictor components based on a cross-validated analysis of hindcasts from CFSv2 climate and land surface fields, and the results are benchmarked against raw CFSv2 forecasts, remapped to intermediate-scale watershed areas. We find that postprocessing affords marginal to moderate gains in skill in many watersheds, raising climate forecast skill above a usability threshold over the four seasons analyzed. In other locations, however, postprocessing fails to improve skill, particularly for extreme events, and can lead to unreliably narrow forecast ranges. This work presents evidence that the statistical postprocessing of climate forecast system outputs has potential to improve forecast skill, but that more thorough study of alternative approaches and predictors may be needed to achieve comprehensively positive outcomes.

The application of sub‐seasonal to seasonal (S2S) predictions for hydropower forecasting

“…Systematic biases exist in the output fields and these biases often grow with lead time, meaning that predictions further into the future are associated with larger biases. Post-processing techniques or calibration can reduce systematic biases and significantly enhance forecast skill, compared with raw output fields (Baker et al, 2020;Manrique-Suñén et al, 2020). Forecast calibration requires time series of historical S2S predictions and corresponding observations to identify and correct systematic biases.…”

Enhancing Ensemble Seasonal Streamflow Forecasts in the Upper Colorado River Basin Using Multi‐Model Climate Forecasts

“…In particular, it is likely that greater climate skill could be harnessed from shorter range forecasts that are more skillful, including weather scale forecasts (which are used at the CBRFC), or S2S forecasts — for example, forecasts for 2–3 weeks lead times, which are not currently used in operations but tend to have higher skill than the NMME one‐ and three‐month forecasts. In separate work by the authors, S2S forecasts from an individual NMME model, Climate Forecast System version 2, were postprocessed and assessed on a HUC‐4 watershed scale, but were not incorporated into this study (Baker and Wood 2020) due to their shorter length of hindcast record.…”

“…Nonetheless, a data‐driven approach to predictor selection, including large‐scale climate system fields such as sea surface temperatures (as in Baker et al. 2020) may lead to improved performance compared to the narrowly focused approach used here. It is also likely that varying the predictors by forecast month — for example, to remove NMME forecasts or other predictors when they do not benefit skill, while keeping them when they do, would provide a benefit.…”

Hybrid forecasting: using statistics and machine learning to integrate predictions from dynamical models

“…Ravuri et al, 2021;Neri et al, 2019), and geographical domains (from point to street-level, single river catchment through to global approaches). Hybrid models have been applied to predict a variety of hydrometeorological variables, including extreme heat and precipitation (Miller et al, 2021;Najafi et al, 2021;Miao et al, 2019;Ma et al, 2022), seasonal climate variables (Golian et al, 2022;Baker et al, 2020), tropical cyclones/hurricanes (Vecchi et al, 2011;Murakami et al, 2016;Kang and Elsner, 2020;Klotzbach et al, 2020), streamflow (Wood and Schaake, 2008;Mendoza et al, 2017;Rasouli et al, 2012;Duan et al, 2020), flooding (Slater and Villarini, 2018), drought (Madadgar et al, 2016;Wu et al, 2021), sea level (Khouakhi et al, 2019), and reservoir levels (Tian et al, 2021), over a range of and predictions have numerous operational and strategic applications, including water resources planning, reservoir inflow management (Tian et al, 2021;Essenfelder et al, 2020), surface water flooding (Rözer et al, 2021), flood risk mitigation, navigation (Meißner et al, 2017), and agricultural crop forecasting (Cao et al, 2022;Slater et al, 2021b). The envisaged dynamical predictors may include various model outputs such as meteorological forecasts with lead times up to 14 days; initialized climate predictions with sub-seasonal to decadal lead times; sub-seasonal runoff predictions, and/or land surface 110 2 Hybrid forecasting Hybrid forecasting encompasses approaches for pre-/post-processing hydroclimate predictions (Section 2.1), and for developing predictive models themselves, including short-term hybrid forecasts (Section 2.2), or sub-seasonal to decadal predictions (Section 2.3), and the integration of ML within parallel and coupled hybrid models (Section 2.4 and Table 3).…”