Abstract:Based on the fish sampling data during fish migration season (March to June, 2006) in the Buyuan River, which is a first order tributary at the downstream of the Lancang River, the fish species diversity was analyzed. With reference to the water quality and hydrological data in the same period, relationship between the environment and the number and weight of fishes were analyzed using artificial neural network (ANN). Results showed that 34 species belonging to 3 orders, 8 families and 22 genera, were collected, among which 8 species were migratory species. Diversity index (H') was at the highest in March, while domination index (Sp) was higher in April and June for the appearance of Mystacoleucus chilopterus and Opsariichthys bidens. There was no significant difference of evenness index (E) among all the sampling months. The lowest water level (LWL) was the most influential factor on the number of fish species, followed by the average runoff velocity (ARV) and average water temperature (AWT). The highest water temperature (HWT) had significant influence on the weight of fish, followed by the total phosphorus and the dissolved oxygen. The simulant values obtained by using ANN were similar with the actual values, which indicated that the ANN method was applicable.
