Poster Session P2.14 Heteroskedasticity and the Weighted Least Square Method in Radar Rainfall Estimate

Monday, 6 August 2007
Halls C & D (Cairns Convention Center)
Wenyao Pu, Nanjing University, Dalian, Liaoning, China; and N. Li, H. Li, Y. He, S. Yu, and Z. Huang

Handout (36.2 kB)

In order to study the relation of echo intensity and precipitation in Dalian and bring into play the effect of Doppler radar and rain gauge for rainfall estimating,the method of Z-I regression is usedBBased on the data of radar and rain gauge from two cumulostratus precipitation processes in 2006,the variety disciplinarian of A and b is analyzedBIt is found there has the heteroskedasticity in the Z-I relation due to its nonlinearity and instability, and the method of weighted least square is proposed for overcoming this problem. Comparing analysis has been done among the method of the weighted least square and others. The research result shows that the instability is restrained in the account processes using the weighted least square method,and the absolute errors of radar rainfall estimation can be minimized while according with practice discipline of Z-I relation.
- Indicates paper has been withdrawn from meeting
- Indicates an Award Winner