Abstract: Amechanism model for simulating the population dynamics of Prorocentrum sigmoides at Dapeng Bay of the South China Sea is established.Among the factors such as temperature, salinity.dissolved oxygen (DO), dissolved inorganic phosphate (DIP), dissolvedinorganic nitrate and nitrite, and PH, the temperature is a limiting factor of P.sigmoidesgrowth.By applying hierachy to the back substitution and introducing auto-regression ordering into the model identification, a simultaneous regression model derived from six auto-regression and non-linear regression models is established.The fitting degree betweensimulated and observed data is 81.7%.

Key words: Prorocentrum sigmoides, Population dynamics, Dapeng, Mechanism model, synthetic aperture radar, support vector machine, QuickBird, object-based, Radarsat