| ET0 forecast on the basis of GA-BP network in high altitude areas of Tibet |
|
View Fulltext View/Add Comment Download reader |
| |
| DOI：10.7606/j.issn.1000-7601.2016.02.34 |
| Key Words: reference crop evapotranspiration (ET0) prediction of ET0 genetic algorithm-back propagation model high altitude areas tibet |
| Author Name | Affiliation | | TANG Peng-cheng | Institute of Water Resources for Pastoral Area of IWHR, Hohhot, Inner Mongolia, 010020, China | | XU Bing | Institute of Water Resources for Pastoral Area of IWHR, Hohhot, Inner Mongolia, 010020, China | | ZHANG Wei-ming | The Tianhe Consulting Limited Company of Hebei, Shijiangzhuang, Hebei 050021, China | | GAO Xiao-yu | College of Water Resources and Civil Engineering, China Agricultural University, Beijing,100083, China | | SONG Yi-fan | Institute of Water Resources for Pastoral Area of IWHR, Hohhot, Inner Mongolia, 010020, China |
|
| Hits: 1449 |
| Download times: 797 |
| Abstract: |
| As the typical climate area of Tibet plateau, Nagqu County (4 450 m above sea level) and Gerze County (4 700 m above sea level) were chosen to build a Genetic Algorithm-Back Propagation (GA-BP) model through the GA-BP network training using data from 1983 to 2012. The ET0 was obtained by the monthly meteorological data from the previous year. When the forecast values for consecutive years between 2010 and 2012 met the threshold limit set, the forecast values would be exported, which could ensure accuracy and stability of the forecast. The results showed that the there was a great linear relationship between the predicted values by GA-BP model and the real values, reaching R2 values of 0.8805, 0.9363 and 0.9167, respectively. The relative error produced by predicted values and real values were all smaller than 0.1, which was less than the threshold. In conclusion, the model can be used to predict the ET0 of different months during crop growth period, and then the inter-monthly water demand of crops can be estimated in the future. It can further provide the basis for future irrigation schedule. For areas lacking meteorological data, based on th
e network model discussed in this article, the ET0 variation in these areas can be simulated within a big time frame by referring to other stations with similar meteorological conditions, which can provide guidance for the irrigation schedule of next year. |
|
|
|