Precipitation forecast based on GRNN model
- Article
- Figures
- Metrics
- Preview PDF
- Reference
- Related
- Cited by
- Materials
Abstract:
Pr ecipitation is the key parameter for water resources manag ement, flood disaster fo recast, and agricultural water usage. Its ac2 curate prediction is of g reat sig nificance for the flood forecast and water resources regulation. In this paper, the precipitation trend in Tieling distr ict was analyzed based on the monthly pr ecipitation data fr om 1960 to 2006. The Mann2Kendall nonparametric method was used to test the significance level on the basis of analysis of climate tendency r ate. The GRNN ( generalized reg ressio n neural netw ork) model w as used in precipitatio n for ecast, and the prediction r esults were analyzed using the error deviat ion and predictio n compariso n fig2 ures. The results showed that prediction in Tieling decreases slow ly in the past 47 years but the tr end hasn. t r eached the sig nifi2 cance lev el ( P > 01 1) . The pr edicted precipit ation w as sim ilar to the actual va lue fo r each mo nth. October has the best predictio n effect w hile Febr ua ry has the least w ith the er ro r o f 31 39% and 191 45%, respectiv ely .
Keywords:
Project Supported:
Clc Number:
Mobile website
