Michael Scheuerer

Modelling daily electricity demand is an essential step to ensure grid stability and to meet society’s needs. Temperature is a key driver of demand, as it not only influences the seasonal variability but also the extremes. Day number is commonly used as a proxy for seasonality and is especially efficient at capturing the lower demand of the summer holiday. This is, however, a static feature and therefore not a sufficient choice when modelling demand in a changing climate. It is therefore of great interest to further investigate how to best resolve the true impact of temperature in demand models.This study quantifies the gain in model performance when utilizing meteorological parameters directly versus using day number only. Furthermore, we evaluate feature engineering strategies to improve the model's ability to leverage the predictive information embedded in temperature. This was done using Generalized Additive Models (GAMs) to model the weather- and calendar-dependent daily electricity demand for nine European countries and assessing different feature combinations.The results demonstrate an overall improvement in model performance when temperature is included in the modelling across all countries. The most significant improvements are seen in the Nordics and France, with up to 51.5% decrease in mean absolute error (MAE) compared to using day number alone. The significance of temperature is most pronounced when assessing model performance on the upper 5th percentile of daily demand, where the reduction in MAE is up to 69.0%. These findings underscore temperature’s critical role in capturing extreme demand events and highlight the need for climate-responsive modelling strategies.

A linear regression model is developed to link anomalies of streamflow to anomalies of precipitation amounts and temperature with the goal of making multi-decadal streamflow projections based on CMIP6 multi-model simulations. Regression coefficients estimated separately for each catchment and each month show physically implausible spatial patterns and indicate issues with overfitting. An alternative approach is therefore explored in which all regression coefficients are estimated simultaneously through a neural network that retains the original linear model structure, but uses embeddings to map each combination of catchment and month to a set of regression coefficients. The model is demonstrated over a set of catchments in Brazil, where the estimated relationships are used to make streamflow projections for the next decades based on CMIP6 multi-model simulations. It yields physically more plausible relationships between streamflow, precipitation amounts, and temperature for our study area than the locally fitted regression models. The resulting projections indicate reduced streamflow over northern, north-eastern, central, and south-eastern Brazil, especially for the austral spring and summer season. The signal is less clear during austral winter. In southern Brazil, an increase in streamflow is expected.

A linear regression model is developed to link anomalies of streamflow to anomalies of precipitation amounts and temperature with the goal of making multi-decadal streamflow projections based on CMIP6 multi-model simulations. Regression coefficients estimated separately for each catchment and each month show physically implausible spatial patterns and indicate issues with overfitting. An alternative approach is therefore explored in which all regression coefficients are estimated simultaneously through a neural network that retains the original linear model structure, but uses embeddings to map each combination of catchment and month to a set of regression coefficients. The model is demonstrated over a set of catchments in Brazil, where the estimated relationships are used to make streamflow projections for the next decades based on CMIP6 multi-model simulations. It yields physically more plausible relationships between streamflow, precipitation amounts, and temperature for our study area than the locally fitted regression models. The resulting projections indicate reduced streamflow over northern, north-eastern, central, and south-eastern Brazil, especially for the austral spring and summer season. The signal is less clear during austral winter. In southern Brazil, an increase in streamflow is expected.

A linear regression model is developed to link anomalies of streamflow to anomalies of precipitation amounts and temperature with the goal of making multi-decadal streamflow projections based on CMIP6 multi-model simulations. Regression coefficients estimated separately for each catchment and each month show physically implausible spatial patterns and indicate issues with overfitting. An alternative approach is therefore explored in which all regression coefficients are estimated simultaneously through a neural network that retains the original linear model structure, but uses embeddings to map each combination of catchment and month to a set of regression coefficients. The model is demonstrated over a set of catchments in Brazil, where the estimated relationships are used to make streamflow projections for the next decades based on CMIP6 multi-model simulations. It yields physically more plausible relationships between streamflow, precipitation amounts, and temperature for our study area than the locally fitted regression models. The resulting projections indicate reduced streamflow over northern, north-eastern, central, and south-eastern Brazil, especially for the austral spring and summer season. The signal is less clear during austral winter. In southern Brazil, an increase in streamflow is expected.

A linear regression model is developed to link anomalies of streamflow to anomalies of precipitation amounts and temperature with the goal of making multi-decadal streamflow projections based on CMIP6 multi-model simulations. Regression coefficients estimated separately for each catchment and each month show physically implausible spatial patterns and indicate issues with overfitting. An alternative approach is therefore explored in which all regression coefficients are estimated simultaneously through a neural network that retains the original linear model structure, but uses embeddings to map each combination of catchment and month to a set of regression coefficients. The model is demonstrated over a set of catchments in Brazil, where the estimated relationships are used to make streamflow projections for the next decades based on CMIP6 multi-model simulations. It yields physically more plausible relationships between streamflow, precipitation amounts, and temperature for our study area than the locally fitted regression models. The resulting projections indicate reduced streamflow over northern, north-eastern, central, and south-eastern Brazil, especially for the austral spring and summer season. The signal is less clear during austral winter. In southern Brazil, an increase in streamflow is expected.

A linear regression model is developed to link anomalies of streamflow to anomalies of precipitation amounts and temperature with the goal of making multi-decadal streamflow projections based on CMIP6 multi-model simulations. Regression coefficients estimated separately for each catchment and each month show physically implausible spatial patterns and indicate issues with overfitting. An alternative approach is therefore explored in which all regression coefficients are estimated simultaneously through a neural network that retains the original linear model structure, but uses embeddings to map each combination of catchment and month to a set of regression coefficients. The model is demonstrated over a set of catchments in Brazil, where the estimated relationships are used to make streamflow projections for the next decades based on CMIP6 multi-model simulations. It yields physically more plausible relationships between streamflow, precipitation amounts, and temperature for our study area than the locally fitted regression models. The resulting projections indicate reduced streamflow over northern, north-eastern, central, and south-eastern Brazil, especially for the austral spring and summer season. The signal is less clear during austral winter. In southern Brazil, an increase in streamflow is expected.

A linear regression model is developed to link anomalies of streamflow to anomalies of precipitation amounts and temperature with the goal of making multi-decadal streamflow projections based on CMIP6 multi-model simulations. Regression coefficients estimated separately for each catchment and each month show physically implausible spatial patterns and indicate issues with overfitting. An alternative approach is therefore explored in which all regression coefficients are estimated simultaneously through a neural network that retains the original linear model structure, but uses embeddings to map each combination of catchment and month to a set of regression coefficients. The model is demonstrated over a set of catchments in Brazil, where the estimated relationships are used to make streamflow projections for the next decades based on CMIP6 multi-model simulations. It yields physically more plausible relationships between streamflow, precipitation amounts, and temperature for our study area than the locally fitted regression models. The resulting projections indicate reduced streamflow over northern, north-eastern, central, and south-eastern Brazil, especially for the austral spring and summer season. The signal is less clear during austral winter. In southern Brazil, an increase in streamflow is expected.

A linear regression model is developed to link anomalies of streamflow to anomalies of precipitation amounts and temperature with the goal of making multi-decadal streamflow projections based on CMIP6 multi-model simulations. Regression coefficients estimated separately for each catchment and each month show physically implausible spatial patterns and indicate issues with overfitting. An alternative approach is therefore explored in which all regression coefficients are estimated simultaneously through a neural network that retains the original linear model structure, but uses embeddings to map each combination of catchment and month to a set of regression coefficients. The model is demonstrated over a set of catchments in Brazil, where the estimated relationships are used to make streamflow projections for the next decades based on CMIP6 multi-model simulations. It yields physically more plausible relationships between streamflow, precipitation amounts, and temperature for our study area than the locally fitted regression models. The resulting projections indicate reduced streamflow over northern, north-eastern, central, and south-eastern Brazil, especially for the austral spring and summer season. The signal is less clear during austral winter. In southern Brazil, an increase in streamflow is expected.

A linear regression model is developed to link anomalies of streamflow to anomalies of precipitation amounts and temperature with the goal of making multi-decadal streamflow projections based on CMIP6 multi-model simulations. Regression coefficients estimated separately for each catchment and each month show physically implausible spatial patterns and indicate issues with overfitting. An alternative approach is therefore explored in which all regression coefficients are estimated simultaneously through a neural network that retains the original linear model structure, but uses embeddings to map each combination of catchment and month to a set of regression coefficients. The model is demonstrated over a set of catchments in Brazil, where the estimated relationships are used to make streamflow projections for the next decades based on CMIP6 multi-model simulations. It yields physically more plausible relationships between streamflow, precipitation amounts, and temperature for our study area than the locally fitted regression models. The resulting projections indicate reduced streamflow over northern, north-eastern, central, and south-eastern Brazil, especially for the austral spring and summer season. The signal is less clear during austral winter. In southern Brazil, an increase in streamflow is expected.

A linear regression model is developed to link anomalies of streamflow to anomalies of precipitation amounts and temperature with the goal of making multi-decadal streamflow projections based on CMIP6 multi-model simulations. Regression coefficients estimated separately for each catchment and each month show physically implausible spatial patterns and indicate issues with overfitting. An alternative approach is therefore explored in which all regression coefficients are estimated simultaneously through a neural network that retains the original linear model structure, but uses embeddings to map each combination of catchment and month to a set of regression coefficients. The model is demonstrated over a set of catchments in Brazil, where the estimated relationships are used to make streamflow projections for the next decades based on CMIP6 multi-model simulations. It yields physically more plausible relationships between streamflow, precipitation amounts, and temperature for our study area than the locally fitted regression models. The resulting projections indicate reduced streamflow over northern, north-eastern, central, and south-eastern Brazil, especially for the austral spring and summer season. The signal is less clear during austral winter. In southern Brazil, an increase in streamflow is expected.