Evaluation of TRMM precipitation data for the estimation of daily stream flows in small watersheds.

---

Raul Antonio Tupac Yupanqui Velez, Raul Antonio Tupac Yupanqui Velez

Thursday 2 july 2015

15:05 - 15:20h at Oceania Foyer (level 0)

Themes: (T) Water resources and hydro informatics (WRHI), (ST) Surface and subsurface flow interactions

Parallel session: 12L. Water resources - Flow interactions

---

Stream flow estimation and prediction are a very important tool for water resources management, however many watersheds are ungauged or present incomplete records for precipitation and discharges, especially in small watersheds at daily level. This work exposes a methodology which contemplates the use of TRMM (Tropical Rainfall Measuring Mission) daily precipitation data in its current version 3B42V7 over small watersheds located in the Northern Peruvian coast, a region affected by climate variability associated with El Niño Phenomenon. This dataset was evaluated into the GR4J lumped model in some scenarios of precipitation data correction and calibration with in-situ data. TRMM correction was based on the wavelet transform, multi-resolution analysis with wavelet and entropy analysis, in function of in-situ precipitation, considering the advantages of this methodology for non-stationary series and a better statistical approximation for point-to-grid downscaling methodologies. This procedure retains the main features of the original signal as the in-situ daily data from which is reconstructed the TRMM as a corrected data. The calibration parameters of GR4J model show a temporal and spatial low variation, allowing its extension and use over incomplete periods or ungauged nearby small watersheds. Finally, stream flows generated with original TRMM present sub estimations with respect to in-situ records in the order of 60% to 70% of Nash´s efficiency at daily level. The corrected TRMM improved this estimation between simulated and in-situ stream flows with 80% of Nash’s efficiency which suggest that TRMM data can be used without correction in some watersheds depending of the climate condition and other factors like elevation, but in the highlands it is necessary to correct TRMM data to improve the application of the GR4J model. These results encourage its use over small watersheds. KEY WORDS: TRMM 3B42V7, precipitation, wavelet transform, GR4J model, stream flow