Statistical Research On Patterns Of Species Richness Of Trees In North America Based On Water-Thermal Energy Hypothesis

The distribution patterns of species richness and their relationship with the environmental factors are the basic problems of biodiversity research.This paper selected parts of North America(the United States and Canada)as the research objects.By collecting the geographical distribution of 679 tree species in North America published on the USGS website and the climate dataset derived from WorldClim website,we focused on the three significant water-thermal energy hypothesis models(Currie productivity hypothesis,PHC;Ecological metabolism hypothesis,MTE;Water-energy dynamics hypothesis,WED).We adopted the least square method to obtain the optimal fitting function for those models.The fitting effects of those three energy hypothesis models at different scales were compared.The conclusions are as follow:(1)For PHC,MTE or WED model,its fitting results are more effective with the increase of scales.(2)The PHC model proposed by Currie in 1987 has a relatively good explanatory effect in all scales.The goodness of model fit from small scale to large scale is 0.676,0.703 and 0.782 respectively.Specially,the results of the median scale including parameter values and explanatory variance are close to the publised paper.(3)The ecological metabolic model proposed by Brown has a good explanatory effect on the large scale,and the explanatory variance can be as high as 0.915471 along with the slope E of the fitting fuction consistent with the hypothesis,which implied a good geographical pattern of species richness.The mesoscale case came the next,in which the model has a 0.69 explanatory variance.However,in the case of small scale,the explanatory effect of the model is not ideal,and the goodness of fit is only 0.22.(4)As for WED model,the model IGM-2 based on the addition of the topographic relief factors has a better fitting effect than IGM-1 at both small and large scales,but no higher than that at mesoscale.For each of the models,the explanatory variance of the model increases monotonically with the increase of the scale,that is,the larger the sampling resolution of the research area,the better the fitting effect of the model.This also shows that the water-energy dynamics model is more suitable for large scale resolution.When the study scale is large,it is appropriate to use the minimum monthly mean temperature to replace the potential evapotranspiration of the coldest month,the value of which were zero in most cases,so that it get closer to the real pattern of species richness.Based on North American trees vegetation distribution information,we compared the fitting effect of the three classic water-thermal energy hypothesis model of species richness.The most applicable scale is obtained for each model and geographical distribution pattern of species richness is studied under different scales,which will provide the verification for the energy hypothesis of the vegetation species diversity pattern.