Development and validation of the CHIRTS-daily quasi-global high-resolution daily temperature data set

Andrew Verdin,Martin Landsfeld,Kathryn Grace,Chris Funk,Cascade Tuholske,Pete Peterson

doi:10.1038/s41597-020-00643-7

Abstract

We present a high-resolution daily temperature data set, CHIRTS-daily, which is derived by merging the monthly Climate Hazards center InfraRed Temperature with Stations climate record with daily temperatures from version 5 of the European Centre for Medium-Range Weather Forecasts Re-Analysis. We demonstrate that remotely sensed temperature estimates may more closely represent true conditions than those that rely on interpolation, especially in regions with sparse in situ data. By leveraging remotely sensed infrared temperature observations, CHIRTS-daily provides estimates of 2-meter air temperature for 1983–2016 with a footprint covering 60°S-70°N. We describe this data set and perform a series of validations using station observations from two prominent climate data sources. The validations indicate high levels of accuracy, with CHIRTS-daily correlations with observations ranging from 0.7 to 0.9, and very good representation of heat wave trends.

Highlights

Background & SummaryThis manuscript is focused on the development and validation of CHIRTS-daily: a high-resolution (0.05° × 0.05°)daily maximum and minimum temperature data series spanning 60°S–70°N
The CHIRTSmax product focuses only on maximum air temperature estimates because minimum temperatures are more difficult to distinguish from the cool cloud-top temperatures observed by satellites
The ERA5 diurnal temperature range (DTR) values are based on the hourly reanalysis simulations, which incorporate physically based land and atmospheric modeling components, and are not affected by potential cloud contamination

Summary

Introduction

Background & SummaryThis manuscript is focused on the development and validation of CHIRTS-daily: a high-resolution (0.05° × 0.05°)daily maximum and minimum temperature data series spanning 60°S–70°N. The CHIRTS-daily data set builds on the Climate Hazards center InfraRed Temperature with Stations Tmax data set (CHIRTSmax1) (Fig. 1a), which leverages remotely sensed infrared land surface emission temperatures and a dense global network of approximately 15,000 Berkeley Earth in situ station observations to provide robust high-resolution (0.05° × 0.05°) estimates of monthly mean maximum 2-meter air temperature (Tmax). For Tmax values, cloud screening can be used to isolate the thermal infrared land surface temperature signal, which correlates well with the monthly maximum 2-meter temperatures observed at weather stations. The ERA5 data set is an advanced reanalysis that uses weather and land models, forced with satellite and in situ observations, to derive a complete suite of physically consistent variables describing many aspects of the Earth.

Objectives

Findings

Methods