探讨运用多元回归分析预报赤潮

对大亚湾澳头水域19972001年赤潮监测资料的统计分析，发现可以以潮汐、风向、天气状况和水温作为影响赤潮发生的重要因子，用参数表征各因的权重，建立多元回归方程，并绘制出赤潮生物的变化趋势图。可根据现场的生物观测资料分析，预报在生物高峰期是否将发生赤潮。该预报方法，可对陆源污染不明显，水质状况较稳定、浮游植物种群丰富的半封闭溺谷湾海水增养殖区的赤潮预报提供参考。

Forecast of red tide with multi-variable regression

By the statistics analysis on monitoring data of red tide in the adjacent waters to Aotou, Daya Bay, water temperature?tide?wind direction and weather can be thought as the major factors affecting red tide. The variables of parameters stand for the factor's weight, multi-variate regression equation can be set up, and the variation trend diagram of red tide organisms is drawn by the applied equation. Based on aquatic biota data observed on the spot, it could be forecasted whether the red tide might occur during red tide organism blooming period. The paper may provide the forecast method of red tide on the basis of the model conducted in the adjacent waters to Aotou, Daya Bay, and on red tide forecast of mariculture farm in semi-closed liman bay with unobstrusively land-based pollution sources, stable water quality and rich phytoplankton population.