Performance of multiple satellite precipitation estimates over a typical arid mountainous area of China : Spatiotemporal patterns and extremes
(2020) In Journal of Hydrometeorology 21(3). p.533-550- Abstract
Precipitation in arid mountainous areas is characterized by low rainfall intensity and large spatial heterogeneity, which challenges satellite-based monitoring by the spaceborne sensors. This study aims to comparatively evaluate the detection ability of spatiotemporal patterns and extremes of rainfall by a range of mainstream satellite precipitation products [TMPA, Climate Hazards Group Infrared Precipitation with Station Data (CHIRPS), and PERSIANN-Climate Data Record (PERSIANN-CDR)] over a typical arid mountainous basin of China, benchmarking against rain gauge data from 2000 to 2015. Results showed that satellite precipitation estimates had relatively low accuracy at the daily scale, while a significant improvement of correlation... (More)
Precipitation in arid mountainous areas is characterized by low rainfall intensity and large spatial heterogeneity, which challenges satellite-based monitoring by the spaceborne sensors. This study aims to comparatively evaluate the detection ability of spatiotemporal patterns and extremes of rainfall by a range of mainstream satellite precipitation products [TMPA, Climate Hazards Group Infrared Precipitation with Station Data (CHIRPS), and PERSIANN-Climate Data Record (PERSIANN-CDR)] over a typical arid mountainous basin of China, benchmarking against rain gauge data from 2000 to 2015. Results showed that satellite precipitation estimates had relatively low accuracy at the daily scale, while a significant improvement of correlation coefficient (CC;>0.6) and a significant reduction of relative root-mean-square error (RRMSE;, 1.0) were found as time scale increases beyond the monthly scale. CHIRPS tended to overestimate the gauge precipitation with positive relative bias (RB), while the negative RBvalues for TMPAand PERSIANN-CDR indicated there was an underestimation. CHIRPS had the most similar spatial pattern and slope trends of the seasonal precipitation and interannual variations of annual precipitation with gauge observations. With the increase in rainfall rates, the probability of detection (POD) and critical success index (CSI) were reduced and the false alarm ratio (FAR) was increased significantly, demonstrating the limited capability for all the three satellite products for detecting heavy rainfall events. CHIRPS showed the best performance in detecting rainfall extremes compared to TMPA and PERSIANN-CDR, evidenced by the largerCSI values and similar extreme rainfall indices obtained from gauge records. This study provides valuable guidance for choosing satellite precipitation products instead of gauge observations for rainfall monitoring (especially rainfall extremes) and agricultural production management over arid mountainous area.
(Less)
- author
- Chen, Cheng ; Li, Zhe ; Song, Yina ; Duan, Zheng LU ; Mo, Kangle ; Wang, Zhiyuan and Chen, Qiuwen
- organization
- publishing date
- 2020
- type
- Contribution to journal
- publication status
- published
- subject
- in
- Journal of Hydrometeorology
- volume
- 21
- issue
- 3
- pages
- 18 pages
- publisher
- American Meteorological Society
- external identifiers
-
- scopus:85083562258
- ISSN
- 1525-755X
- DOI
- 10.1175/JHM-D-19-0167.1
- language
- English
- LU publication?
- yes
- id
- 448fda09-1c5e-41fd-8efd-8cf1606564c8
- date added to LUP
- 2021-01-15 09:40:14
- date last changed
- 2022-04-26 23:35:00
@article{448fda09-1c5e-41fd-8efd-8cf1606564c8,
abstract = {{<p>Precipitation in arid mountainous areas is characterized by low rainfall intensity and large spatial heterogeneity, which challenges satellite-based monitoring by the spaceborne sensors. This study aims to comparatively evaluate the detection ability of spatiotemporal patterns and extremes of rainfall by a range of mainstream satellite precipitation products [TMPA, Climate Hazards Group Infrared Precipitation with Station Data (CHIRPS), and PERSIANN-Climate Data Record (PERSIANN-CDR)] over a typical arid mountainous basin of China, benchmarking against rain gauge data from 2000 to 2015. Results showed that satellite precipitation estimates had relatively low accuracy at the daily scale, while a significant improvement of correlation coefficient (CC;>0.6) and a significant reduction of relative root-mean-square error (RRMSE;, 1.0) were found as time scale increases beyond the monthly scale. CHIRPS tended to overestimate the gauge precipitation with positive relative bias (RB), while the negative RBvalues for TMPAand PERSIANN-CDR indicated there was an underestimation. CHIRPS had the most similar spatial pattern and slope trends of the seasonal precipitation and interannual variations of annual precipitation with gauge observations. With the increase in rainfall rates, the probability of detection (POD) and critical success index (CSI) were reduced and the false alarm ratio (FAR) was increased significantly, demonstrating the limited capability for all the three satellite products for detecting heavy rainfall events. CHIRPS showed the best performance in detecting rainfall extremes compared to TMPA and PERSIANN-CDR, evidenced by the largerCSI values and similar extreme rainfall indices obtained from gauge records. This study provides valuable guidance for choosing satellite precipitation products instead of gauge observations for rainfall monitoring (especially rainfall extremes) and agricultural production management over arid mountainous area.</p>}},
author = {{Chen, Cheng and Li, Zhe and Song, Yina and Duan, Zheng and Mo, Kangle and Wang, Zhiyuan and Chen, Qiuwen}},
issn = {{1525-755X}},
language = {{eng}},
number = {{3}},
pages = {{533--550}},
publisher = {{American Meteorological Society}},
series = {{Journal of Hydrometeorology}},
title = {{Performance of multiple satellite precipitation estimates over a typical arid mountainous area of China : Spatiotemporal patterns and extremes}},
url = {{http://dx.doi.org/10.1175/JHM-D-19-0167.1}},
doi = {{10.1175/JHM-D-19-0167.1}},
volume = {{21}},
year = {{2020}},
}