Empirical Model Construction Of Area And Mass Dry For A Single Leaf Of Main Broad-leaved Species In Northeast China

The empirical model method can conveniently and quickly measure the leaf area(LA)and leaf dry mass(LM)of broad-leaved plants under non-destructive conditions,but an empirical model that can efficiently measure the LA and LM of different leaf-type broad-leaved plants at different leaf growth periods has not been proposed yet.In this experiment,twelve broadleaf species of Acer pictum subsp.mono,Acer ukurunduense,Tilia amurensis,Acer tegmentosum,Betula platyphylla,Corylus mandshurica,Betula costata Ulmus laciniata,Syringa reticulata subsp.amurensis,Ulmus davidiana var.japonica,Lonicera japonica,Fraxinus mandschurica in northeast China were selected as the research object,to separately measure the leaves length(L),leaves width(W),leaf thickness(T),LA and LM at different stages(leaf growth period,leaf stability period and leaf fall period).Then we explored the changes of these leaf traits;Using L,W,T and their combinations,the separate model for predicting LA and LM of a single leaf of an specie in different periods was constructed;To simplify the cumbersome procedure of building different empirical models due to different plant,all plants were classified and combined based on the leaf length-to-width ratio(L/W),than combined empirical models were constructed suitable for predicting LA and LM of a single leaf of plants with similar leaf patterns in different leaf growth period.At the same time,the empirical models were evaluated and tested.(1)There was a certain variation in leaf traits in broad-leaved plants,and the coefficient of variation from large to small was LM>LA>T>W>L>L/W,and the coefficient of variation of L/W was 9%～21%.Therefore,L/W was often used as a stable structural parameter to identify and classify plants.(2)The best independent variable of leaf area individual model was the combination of L and W(including LW and Lb Wc),and most of the optimal models are power functions,and the relative error of LA predicted by separate models in different growth stages of broad-leaved plants was 4%～15%;The optimal models of the combined models were LA=aLbWc(a、b and c are the model coefficients),and the relative error of predicting LA was 4%～17%,which was only 3%higher than that of the separate models,indicating that the combined model can quickly and efficiently determine LA of the same leaf type broad-leaved plants.(3)The optimal independent variables and optimal models of separate model of LM were greatly affected by tree species and growing period,and the selections were diverse,and the relative error of predicting LM in different growth periods of broad-leaved leaves was 8%-28%;The optimal models of the combined models were power function and optimal independent variables were LW and Lb Wc,it was suitable for predicting LM of leaf growth period and leaf stability period,with a relative error of 10%～29%.(4)Time variation have significant impact on the empirical models of LA and LM,and different empirical models need to be built for different growth period of trees.The results of this study can provide technical support for rapid and efficient determination of the dynamic changes of LA and LM of a single leaf of broad-leaved plants under non-destructive conditions.