A High-Resolution Quantitative Precipitation Estimate over Alaska through Kriging-Based Merging of Rain Gauges and Short-Range Regional Precipitation Forecasts

• Published in: Journal of atmospheric and oceanic technology Vol. 39; no. 6; pp. 739 - 753
• Main Authors: Hoover, Brett T., Otkin, Jason A., Petrescu, Eugene M., Niebuhr, Emily
• Format: Journal Article
• Language: English
• Published: Boston American Meteorological Society 01.06.2022
• Subjects: Dynamic models; Error analysis; Estimates; Forecasting data; Gauges; High resolution; Numerical prediction; Numerical weather forecasting; Precipitation; Precipitation estimation; Precipitation forecasting; Prediction models; Rain; Rain gauges; Rainfall; Remote sensing; Resolution; Statistical methods; Uncertainty analysis; Variance analysis; Weather forecasting; Work platforms; United States > US; Alaska
• ISSN: 0739-0572; 1520-0426
• DOI: 10.1175/JTECH-D-21-0132.1

Abstract A method is presented to generate quantitative precipitation estimates over Alaska using kriging to merge sparse, unevenly distributed rain gauge observations with quantitative precipitation forecasts from a three-member ensemble of high-resolution numerical weather prediction models. The estimated error variance of the analysis is computed by starting with the estimated error variance from kriging and then refining the variance in k -fold cross validation by an empirically derived inflation factor. The method combines dynamical model forecast information with observational data to deliver a best linear unbiased estimate of precipitation, along with an analysis uncertainty estimate, that provides a much-needed precipitation analysis in a region where sparse in situ observations, poor coverage by remote sensing platforms, and complex terrain introduce large uncertainties that need to be quantified. For 6-hourly accumulation estimates produced four times daily from 1 August 2019 to 31 July 2020, three analysis configurations are tested to measure the value added by including model forecast data and how those data are best utilized in the analysis. Several directions for further improvement and validation of the analysis product are provided.