Keywords

Keywords: Climate change impact assessment, inverse approach, stochastic rainfall

Start Date

26-6-2018 3:40 PM

End Date

26-6-2018 5:00 PM

Abstract

foreSIGHT (Systems Insights from Generation of Hydroclimatic Timeseries) is a new R package for performing climate impact assessments using a scenario-neutral approach. These approaches ‘stress test’ a modelled system using a broad range of hydroclimate scenarios, and thus rely on the availability of suitable sets of plausible hydroclimate variable time series. However, there are significant technical challenges in creating the necessary scenarios to explore system sensitivity, and computational overheads in simulating and visualizing system performance in response to large sets of hydroclimate scenarios.

In answer to these challenges foreSIGHT can generate perturbed time series using a range of approaches (e.g. scaling of observed time series, stochastic simulation of perturbed time series via an inverse approach), incorporating a number of stochastic models to generate different hydroclimate variables on a daily basis (e.g. precipitation, temperature, evapotranspiration). The software utilises formal optimisation techniques with the stochastic models to create desired time series which, crucially, allows a variety of different hydroclimate variable properties to be perturbed (e.g. means, percentiles, persistence). This allows for a greater exploration of system sensitivity.

The software allows for the integration of existing system models, both internally in R and externally, and provides a suite of visualization options for the results of a scenario-neutral analysis. This is demonstrated on a simplified water supply system with a climate dependent demand model, where different management configurations are evaluated, and visualised with climate model projections to add context. The necessary optimisation techniques to create perturbed time series are demonstrated using five sites around Australia.

Stream and Session

Stream F: System Identification Approaches for Complex Environmental Systems

F3: Modelling and Decision Making Under Uncertainty

COinS
 
Jun 26th, 3:40 PM Jun 26th, 5:00 PM

An R tool for scenario-neutral climate impact analysis of water resource systems

foreSIGHT (Systems Insights from Generation of Hydroclimatic Timeseries) is a new R package for performing climate impact assessments using a scenario-neutral approach. These approaches ‘stress test’ a modelled system using a broad range of hydroclimate scenarios, and thus rely on the availability of suitable sets of plausible hydroclimate variable time series. However, there are significant technical challenges in creating the necessary scenarios to explore system sensitivity, and computational overheads in simulating and visualizing system performance in response to large sets of hydroclimate scenarios.

In answer to these challenges foreSIGHT can generate perturbed time series using a range of approaches (e.g. scaling of observed time series, stochastic simulation of perturbed time series via an inverse approach), incorporating a number of stochastic models to generate different hydroclimate variables on a daily basis (e.g. precipitation, temperature, evapotranspiration). The software utilises formal optimisation techniques with the stochastic models to create desired time series which, crucially, allows a variety of different hydroclimate variable properties to be perturbed (e.g. means, percentiles, persistence). This allows for a greater exploration of system sensitivity.

The software allows for the integration of existing system models, both internally in R and externally, and provides a suite of visualization options for the results of a scenario-neutral analysis. This is demonstrated on a simplified water supply system with a climate dependent demand model, where different management configurations are evaluated, and visualised with climate model projections to add context. The necessary optimisation techniques to create perturbed time series are demonstrated using five sites around Australia.