Using Calibrated Rainfall Forecasts and Observed Rainfall to Produce Probabilistic Meteorological Drought Forecasts

Most existing drought forecast systems rely only on observed or forecast rainfall, losing valuable context gained from considering both. The lack of a direct link between observed and forecast rainfall reduces the physical consistency of a system, motivating the development of a methodology that can...

Full description

Saved in:
Bibliographic Details
Published in:Climate (Basel) 2024-04, Vol.12 (4), p.54
Main Authors: Chua, Zhi-Weng, Kuleshov, Yuriy, Bhardwaj, Jessica
Format: Article
Language:English
Subjects:
Calibration
Climate models
Climatology
Datasets
Drought
Drought forecasting
drought monitoring
Ensemble forecasting
Environmental aspects
Measurement
Methodology
Methods
Monthly forecasts
Precipitation
Probability forecasting
Rain
Rain and rainfall
Rainfall
rainfall analysis
Rainfall forecasting
Seasonal forecasting
Statistical analysis
Variables
Water shortages
Weather forecasting
Citations: Items that this one cites
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Most existing drought forecast systems rely only on observed or forecast rainfall, losing valuable context gained from considering both. The lack of a direct link between observed and forecast rainfall reduces the physical consistency of a system, motivating the development of a methodology that can directly link the two. The methodology developed in this study allows the comparison of the calibrated ensemble forecasts of rainfall totals from a dynamical climate model to observed rainfall deficiencies from a gridded rainfall analysis. The methodology is used to create a probabilistic product that forecasts the chance of entering meteorological drought, with lead times of one month (monthly forecast) and three months (seasonal forecast). Existing deficiency areas are included to facilitate analysis of how these areas are forecast to change. The performance of the developed methodology was verified using Percent Correct (PC), Brier Score (BS), and Relative Operating Characteristic (ROC) statistics. Analysis of the forecast plots was also completed visually. Forecast performance for areas with existing deficiencies as well as for non-deficiency areas was promising (PC rates of >79% and >97%, respectively). Although PC rates for observed deficiencies were low across most months, the mean forecast probability for these areas was 36%, indicating the system had value and outperformed climatology. A calibrated, coupled product like the one scoped in this study has not been explored and we note that it could be an invaluable tool for quantifying meteorological drought onset and persistence in Australia.
ISSN:2225-1154
2225-1154
DOI:10.3390/cli12040054