A Hidden Climate Indices Modeling Framework for Multivariable Space‐Time Data
Résumé
Resources and risk management for environmental and technological systems often requires information on multiple dependent and spatially distributed variables whose properties vary in time under the influence of climate variability and change. For instance, wildfire depends on the intensity, spatial extent, and possible co-occurrence of droughts, heat waves, and high winds (Barbero et al., 2014; Sharples et al., 2016). Each of these components may be affected by some climate-related trend (e.g., Perkins-Kirkpatrick et al., 2016, for heat waves) or multidecadal variability (e.g., Kiem et al., 2016, for droughts). As another example, renewable energy production depends on the space-time variability of multiple variables such as wind, solar radiation, and river streamflow (Engeland et al., 2017; François et al., 2014). Quantitative risk assessment for such systems therefore relies on a probabilistic model for several spatially distributed variables. In particular, this model should adequately describe spatial and intervariable dependencies in order to obtain reliable risk estimates. It should also account for possible climate-related temporal variability or trend. Finally, this model should be flexible enough to accommodate the complicating factors frequently
Origine | Publication financée par une institution |
---|