University of Sussex
Browse

File(s) not publicly available

The development of composite algorithms to estimate global rainfall using passive microwave data from DMSP SSM/I

journal contribution
posted on 2023-06-07, 17:03 authored by Dominic KnivetonDominic Kniveton, E C Barrett, J C Dodge
Considerable variation in the performance of passive microwave global rainfall algorithms, both spatially and temporally, was revealed by the first WetNet Precipitation Intercomparison Project, PIP-1, with no one algorithm achieving the best results, in all locations, and all the time. In this paper a Composite Algorithm Procedure is described for the Special Sensor Microwave/Imager (SSM/I) algorithms submitted to PIP-1, that attempts through combining the best algorithm results from different regions of the globe to achieve better overall global rainfall estimates than are possible from any individual algorithm alone. The Composite Algorithm Procedure (CAP) involves the segmentation of the globe into homogeneous regions, the production of validation statistics for the various algorithm results in the different regions, and the identification of combinations of algorithms which perform best globally. The segmentations were based on aspects of the spatial and temporal variability of rainfall, or the microwave properties of the surfaces of the Earth. Initial results for the Composite Algorithm Procedure are presented for a sample month (October 1987): these confirm that improved global rainfall products can be produced in this way. Code detailing a selected Composite Algorithm based on the segmentation method of the microwave properties of the Earth has been supplied to the WetNet Support Group at the Marshall Space Flight Center, Huntsville, Alabama, for experimental, regular production of global rainfall data sets on a near real-time basis.

History

Publication status

  • Published

Journal

International Journal of Remote Sensing

ISSN

01431161

Publisher

International Journal of Remote Sensing

Issue

3

Volume

18

Page range

517-534

ISBN

0143-1161

Department affiliated with

  • Geography Publications

Full text available

  • No

Peer reviewed?

  • Yes

Legacy Posted Date

2012-02-06

Usage metrics

    University of Sussex (Publications)

    Categories

    No categories selected

    Exports

    RefWorks
    BibTeX
    Ref. manager
    Endnote
    DataCite
    NLM
    DC