Evaluation of GPM IMERG Rainfall Estimates under Different Rain Gauge Densities in High Elevation Tropical Andes
Abstract
Satellite-based estimates of rainfall are now frequently being used to complement scarce networks of gauges particularly in remote and mountainous regions. Understanding the associated uncertainties is important before the data are applied, but understanding is often limited due to low density of ground gauges and associated scale-related errors. This is of particular importance in mountain regions, which are characterised by complex topography and highly variable rainfall patterns over short spatial scales that are generally smaller than the resolution of the satellite products.
This paper utilises a unique dataset containing 15 rain gauges at elevations ranging from 3,800 to 4,600 m above sea level to evaluate GPM IMERG V5 performance in the tropical Andes of Peru. The 15 gauges represent one IMERG grid so provide a rare evaluation opportunity in this geographical context. IMERG performance was evaluated from two main perspectives. First, both IMERG "Final" and "Near Real Time" products are evaluated at the native resolution (30 minutes) using a combination of performance metrics. Error dependencies on rainfall characteristics and source instrument / processing algorithm are also investigated. Second, the impact of gauge density on perceived 'performance' is examined at different temporal scales (30 minutes to 24 hours) by evaluating different gauge combinations. Our findings show that IMERG performance is limited at the 30-minute resolution predominantly due to significant under-detection of events, while Near Real Time products have better skill than IMERG-Final at rainfall intensity estimation. Performance is strongly influenced by the source instrument and especially by rainfall intensity and variability. We conclude by highlighting the influence of gauge density on performance and provide recommendations on how this problem might be accounted for by IMERG end users.- Publication:
-
AGU Fall Meeting Abstracts
- Pub Date:
- December 2019
- Bibcode:
- 2019AGUFM.H31P1989B
- Keywords:
-
- 3354 Precipitation;
- ATMOSPHERIC PROCESSES;
- 3360 Remote sensing;
- ATMOSPHERIC PROCESSES;
- 1655 Water cycles;
- GLOBAL CHANGE;
- 1840 Hydrometeorology;
- HYDROLOGY