Calibration methods for spatial Data

TL;DR

This paper discusses calibration methods for spatial environmental data, focusing on wind speeds, to improve the accuracy of risk maps by aligning simulated data with observed extreme values.

Contribution

It introduces a novel conditional quantile matching calibration method tailored for non-linear, non-Gaussian extreme wind data, enhancing data reliability for risk analysis.

Findings

Calibration improves the match between simulated and observed wind extremes.

The proposed method effectively adjusts the tail behavior of simulated data.

Enhanced data calibration leads to more accurate risk mapping.

Abstract

In an environmental framework, extreme values of certain spatio-temporal processes, for example wind speeds, are the main cause of severe damage in property, such as electrical networks, transport and agricultural infrastructures. Therefore, availability of accurate data on such processes is highly important in risk analysis, and in particular in producing probability maps showing the spatial distribution of damage risks. Typically, as is the case of wind speeds, data are available at few stations with many missing observations and consequently simulated data are often used to augment information, due to simulated environmental data being available at high spatial and temporal resolutions. However, simulated data often mismatch observed data, particularly at tails, therefore calibrating and bringing it in line with observed data may offer practitioners more reliable and richer data sources. Although the calibration methods that we describe in this manuscript may equally apply to other environmental variables, we describe the methods specifically with reference to wind data and its consequences. Since most damages are caused by extreme winds, it is particularly important to calibrate the right tail of simulated data based on observations. Response relationships between the extremes of simulated and observed data are by nature highly non-linear and non-Gaussian, therefore data fusion techniques available for spatial data may not be adequate for this purpose. After giving a brief description of standard calibration and data fusion methods to update simulated data based on the observed data, we propose and describe in detail a specific conditional quantile matching calibration method and show how our wind speed data can be calibrated using this method.