The goal is to develop a global scale daily precipitation product based
on existing multi-satellite products and bias-corrected precipitation
gauge analyses for
ONC and OFM
applications. The objective is to improve
the GPCP (Global Precipitation Climatology Project)-1DD global product
(Huffman et al, 2001), verified over Europe
by Rubel et al. (2002), with a better calibration making use of
daily bias-corrected rain gauge analyses based on about 6000 synoptic
stations.
IMP and
EARS assess
the precipitation amount from geo-stationnary sensors images
using various approaches.
Combined daily satellite gauge
precipitation product for July 2000.
IMP uses the GPCP-1DD multi-satellite
product derived from geo-stationary sensors data in the 40° north-south
belt, and TOVS data from the polar orbiting NOAA 12 and NOAA 14. This product
is calibrated with monthly precipitation data being near-real-time available
based upon SYNOP and CLIMAT reports from GPCC (Global Precipitation
Climatology Center). Currently, there does not exist any operational global
daily precipitation product, which is based on bias-corrected gauge analyses.
IMP has collected global daily
rain gauge data for the period 1997-2003, and to correct them for
systematic measurement errors. The bias correction of the
ground-based precipitation measurements is needed, because of the
under-catch of operational rain gauges, which is on the order of
5-30 percent on average. Then, the rain gauge data is interpolated
to a global regular 1° longitude/latitude grid. GPCP-1DD satellite
product and rain-gauge measurements are co-kriged to estimate the
final CSP precipitation product. Based on the first results of the FP5
demonstration project
ELDAS, a combined gauge-satellite data set for
the period 1997-present has been compiled and verified over selected
regions of the globe, mainly located in Europe. This includes the
collection of dense precipitation measurements (non-synoptic) for
verification purposes as well as the calculation of verification scores
to demonstrate the expected improvements.
Monthly precipitation from METEOSAT images,
February 2003.
EARS provides the rainfall product
generated by the EWBMS
system. The retrieval algorithm is based upon
the discrimination of cloud top levels in hourly satellite imagery,
using temperature thresholds. For each level the cloud duration
is counted hourly during a period of 10 days. Rain gauge data are
obtained daily from the WMO-GTS system. A local regression based on
the observations on a given rainfall station and on the closest 11
surrounding stations is applied. The regression coefficients and
the residual are subsequently interpolated between the rainfall
stations for each pixel. Hereafter, the rainfall is determined on a
pixel basis.
The IMP and EARS methodologies are similar but not exactly the same. An
inter-comparison of the product will be realized to improve the
reliability of each product. One methodology could be to compare
rainfall products with a validation data set generated by means of
Jackknifing.
References
Huffman, G.J., R.F. Adler, M.M. Morrissey, S. Curtis, R. Joyce, B.
McGavock, and J. Susskind, Global precipitation at one-degree
daily resolution from multi-satellite observations, J. Hydrometeor.,
2, 36-50, 2001.
Rubel, F., P. Skomorowski, and B. Rudolf, Verification scores for
the operational GPCP-1DD product over the European Alps, Meteorol. Z.,
11, 367-370, 2002.
