Identifying weather regimes for regional‐scale stochastic weather generators

Abstract

Weather regime based stochastic weather generators (WR‐SWGs) have recently been proposed as a tool to better understand multi‐sector vulnerability to deeply uncertain climate change. WR‐SWGs can distinguish and simulate different types of climate change that have varying degrees of uncertainty in future projections, including thermodynamic changes (e.g., rising temperatures, Clausius‐Clapeyron scaling of extreme precipitation) and dynamic changes (e.g., shifting circulation and storm tracks). These models require the accurate identification of WRs that are representative of both historical and plausible future patterns of atmospheric circulation, while preserving the complex space–time variability of weather processes. This study proposes a novel framework to identify such WRs based on WR‐SWG performance over a broad geographic area and applies this framework to a case study in California. We test two components of WR‐SWG design, including the method used for WR identification (Hidden Markov Models (HMMs) vs. K‐means clustering) and the number of WRs. For different combinations of these components, we assess performance of a multi‐site WR‐SWG using 14 metrics across 13 major California river basins during the cold season. Results show that performance is best using a small number of WRs (4–5) identified using an HMM. We then juxtapose the number of WRs selected based on WR‐SWG performance against the number of regimes identified using metastability analysis of atmospheric fields. Results show strong agreement in the number of regimes between the two approaches, suggesting that the use of metastable regimes could inform WR‐SWG design. We conclude with a discussion of the potential to expand this framework for additional WR‐SWG design parameters and spatial scales.

Understanding the Contributions of Paleo‐Informed Natural Variability and Climate Changes to Hydroclimate Extremes in the San Joaquin Valley of California

“…Ensembles of plausible future climate are generated using extensions of the WR‐based stochastic weather generator presented in Najibi et al. (2021) and Steinschneider et al. (2019) to incorporate paleo‐reconstructions of WR dynamics.…”

“…All together, we develop 25 different scenarios of climate change (five temperature scenarios and five scaling scenarios), with each scenario containing 50 ensemble members (i.e., 50 stochastic 600-year time series of precipitation and temperature), in addition to a baseline ensemble with no changes imposed. Technical details on the quantile mapping procedure, and other details of the stochastic weather generator, are provided in Najibi et al (2021) and Steinschneider et al (2019).…”

“…Ensembles of plausible future climate are generated using extensions of the WR-based stochastic weather generator presented in Najibi et al (2021) and Steinschneider et al (2019) to incorporate paleo-reconstructions of WR dynamics. The generator is comprised of a three-step hierarchical structure (Figure 3): (a) identification components are used to force a non-homogeneous hidden Markov model (NHMM), whereby WR states are modeled as a first-order Markov chain with a non-stationary transition probability matrix conditioned on the reconstructed PC WR from Gupta et al (2023a).…”

“…To answer these questions, we contribute a framework for creating a regionally consistent ensemble of plausible daily future climate and streamflow scenarios that represent natural climate variability captured in a network of tree-ring chronologies, and then embed anthropogenic climate change trends within those scenarios. A key contribution of this study is a method to utilize 600 years of paleo-informed reconstructions of annual weather regimes (WRs; Gupta et al, 2023a) to force a weather-regime based daily-scale stochastic weather generator (Najibi et al, 2021;Steinschneider et al, 2019), which we utilize to force hydrologic models of five watersheds in the San Joaquin River basin. This novel component of the framework allows for the generation of regional precipitation and temperature that represents the substantial natural atmospheric variability that characterizes the San Joaquin region, addressing the need to integrate such variability into data sets used to inform robust water resources planning and management (Lehner & Deser, 2023).…”

Precipitation Scaling With Temperature in the Northeast US: Variations by Weather Regime, Season, and Precipitation Intensity

“…To identify WRs, we followed Najibi et al. (2021) and fit first‐order Hidden Markov Models (HMMs) to the first J principal components of GPHAs, with J selected to explain a large majority (>90%) of variance. The result partitioned each day in the record into one of K separate WRs (or states) in that season.…”