The Research Of The Prediction Of The Natural Rhizophora Stylosa Individual Tree’s Viviparous Propagule Yield

Natural reproduction and forestation are widely used in the production of Rhizophora stylosa in China,so viviparous propagator is an extremely important basic material for the forestation of Rhizophora stylosa.The production quantity and quality of viviparous propagator of Rhizophora stylosa will directly affect the quality of the forestation.Prediction of mangrove viviparous propagule yield is an important basic work f or scientific organization of mangrove production,afforestation and mangrove ecosystem management.This paper is based on five cities which own natural Rhizophora stylosa community,and the data is from two different years of GaoQiao Mangrove Preserve,XvWen County in Guangdong Province,and CaiQiao Mangrove Preserve of LinGao County,DongZhai Harbor Mangrove Preserve,QingLan Harbor Mangrove Preserve of Hainan Province.Finding the main factors through a way of Pearson correlation analysis,then creating the predicted model through a way of multiple stepwise regression analysis.Finally,selecting the best predicted model among the models created before through ways of comparing residual plot attribution and variance analysis and independent-sample t test.The major conclusions of this paper are:(1)Different areas have different correlative factors,the individual tree’s yield of GaoQiao Mangrove Preserve is correlative to height,ground diameter,crown width and the thickness of lateral branches;the individual tree’s yield of XvWen County is correlative to height,ground diameter,crown width,the shape index of seed and height under branch;the individual tree’s yield of DongZhai Harbor Mangrove Preserve is correlative to amount of lateral branch,ground diameter,crown width and the thickness of lateral branches;the individual tree’s yield of QingLan Harbor Mangrove Preserve is correlative to the shape index of seed,ground diameter,crown width and the thickness of lateral branches;the individual tree’s yield of CaiQiao Mangrove Preserve is correlative to height,crown width and the height of lateral branches.After making all the data into an integrated whole,then we find that the comprehensive individual tree’s yield is correlative to the shape index of seed,ground diameter and the thickness of lateral branches.Then taking the site factors into consideration,we can get the conclusion that the comprehensive data which added site factors is correlative to total nitrogen,total phosphorus,total potassium,hydrolysable nitrogen and the shape index of seed,ground diameter and the thickness of lateral branches.(2)Creating predicted models through a way of multiple stepwise regression analysis,and testing these models through a way of independent-sample t test,finally we find that all the subdivision predicted models has a good precision,which are all higher than 80%,and the comprehensive predicted model which is created by site factors and measurement factors also has a precision of 84.7%,even though the comprehensive predicted models which is based on the measurement factors has a lower precision of 78%,it’s coefficient of determination is 0.794,all the data above can be a proof which can prove the models are feasible in a way.(3)Choosing the latest data to practical test,according to the result,selecting the best predicted model,it’s from the practical test.After comparing the results of residual plot attribution and variance analysis and independent-sample t test,finally we get the conclusion that the comprehensive predicted model which is based on site factors and measurement factors is the best one,because it can overcome regional heterogeneity when the level of the error and precision is in a same condition.