中文摘要：

基于Elman网络基本原理，建立了布洋1号表层岩溶泉Elman神经网络模型，并以2005年8月2日开始的一次衰减过程为例，详述了Elman神经网络模型应用分析过程，给出了在无降雨影响下该次衰减过程具有的总排泄量及最佳的储水时间。经检验，该模型预测精度较高，为布洋1号表层岩溶泉水资源的科学利用提供了依据，同时也为Elman网络技术在表层岩溶泉动态系统的其它领域应用提供了借鉴。

英文摘要：

Taken No. 1 epikarst spring at Buyang as an example, a neural network model has been constructed based on the Elman network principle. Analyses made with this model show that the precision of both the simulation and prediction is high, so this model can be used to forecast the regime of epikarst spring. The attenuation process analyses made by the combination of this model and attenuation theory on August in 2005 tell us the total discharge and the best time for saving the spring water when there is no rain. The conclusion can be used to develop this spring water better. And the application of Elman neural network is also a reference for the other epikarst spring dynamic system.