Site Quality Index Of Arbor Forest Based On DBH Ultimate Growth

Estimate of forest site quality was one of the basic work in forestry production, which could provide scientific forest management measures and predicted the forest harvest precisely. It played a vital role in forestation, forest cultivation and management. This study used asymptote of growth model as a new index of site quality, which stood for the theoretically maximal DBH value. The permanent sampling plot data of Zhejiang province was used for modeling, we built the DBH growth model by two observations in this data, site quality also could be estimated by using the two observations including the mean DBH and periodic increment of DBH. In this model we took stand density and competition index into consideration. While comparing with traditional site quality index, this model overcame some disadvantages in the former like relying on stand age and pure forest stand.This study estimated the new site quality index called M index by using Richards growth model and Schmacher growth model, in the study a new method which calls PSLM was used to estimate parameters, we finally drew a conclusion that variance analysis of M index in Schmacher model showed a higher precision than the result of Richards, which proved that Schmacher model performed well and we chose Schmacher model as the best model. A equationDytD=Dwas assumed to made the M model, The model of pinaceae tree was0.80255.16(1-)eyD CIM =D With the F test value of M index is 7.861. Similarly, we could get the model0.299521.94(1-)eyD CIM =D and F test value 9.614 in taxodiaceae tree, we could also get the model1.21652.7(1-)eyD CIM =D and F test value 4.308 in broadleaf tree. We found that F value had significant difference in all three kind of tree, which meant the new site quality M index is valuable and it had significant result. Besides,we analysed correlation among the three M index, we also found that M index of pineaceae and taxodiaceae tree had best correlation with a high R2 value 0.6029, M index of pineaceae and broadleaf tree had little correlation with a low R2 value 0.2721, M index of broadleaf and taxodiaceae tree almost had no correlation, which was in conformity with the real situation.Least Square Method(LSM) might result in some bad estimate of parameter while dealing with bad modeling data, this study provided a new parameter estimate method – PSLM, which was short for Non-linear Parameter Square Limit Method, this method used a limited parameter k to restrain the parameter so that the model could get better parameters. In the study we found that M index and PLSM was subjective method and more study should be done to improve it.