Responses Of Camphor Forest Soil Respiration To Nitrogen Addition Based On GA-BP Network

The current study on responses of soil respiration to nitrogen addition was observed.Restricted by natural conditions,it can’t focus on the studies of soil respiration to the continuous variation of nitrogen response in some certain conditions.A simulated nitrogen deposition experiment has been conducted in Camphor forest,which located in Hunan Forest Botanical Garden,subtropical China during May 2010 and June 2012.Soil respiration rate was measured twice a month under four levels of N treatments.By spraying NH4NO3 aqueous solution,this research set control(C,no N added),low nitrogen(L,5 g Nm-2 a-1),middle nitrogen(M,15 g N m-2 a-1),and high nitrogen(H,30 g N m12 a。1)as four kinds of processing level,accompanying by using GA-BP artificial neural network to establish forest soil respiration of nitrogen addition model,and the final datas were simulated by the responding surface method.The total results was demonstrated below:1.GA-BP artificial neural network can satisfactory simulate the response of soil respiration to different nitrogen application levels of Cinnamomum camphora in subtropical area.The results of the model simulation were continuous and complete,which explained the soil respiration process(0.70-0.74)under the interaction of nitrogen application level,soil temperature and soil moisture.2.Firstly,each date of the matrix by using soil temperature as the x-axis and soil moisture as the y-axis was input to the model simulation.Then,setting the analog value as a z-axis to the graph,the response surface map and the corresponding contour map of soil respiration was obtained,corresponding to soil temperature and soil moisture.The response surface map and rate of change chart described the continuous and complete transformation,of which soil respiration was responsive to soil temperature and soil moisture under different treatment levels,intuitively.As a whole,the soil respiration rate was suppressed by nitrogen application.The soil respiration rate under H treatment was increased when compared to the soil respiration rate under L、M treatment,but lower than that of C treatment.The lower soil respiration rate mainly occurs in the area of lower soil temperature,and the minimum value appears in the area where the soil temperature and soil humidity are lower.However,the higher soil respiration rate mainly occurs in the area where the soil temperature and soil humidity are higher.Overall,the differences of soil respiration in different humidity ranges was reduced upon the nitrogen application rate increased.The ratio of total soil respiration,soil respiration of low soil moisture area,soil respiration of high soil moisture area under different treatment were as follows:1:0.83:1.29(C treatment),1:0.84:1.18(L treatment),1:0.95:1.08(M treatment)and 1:1.04:0.96(H treatment).It shows that the soil respiration in different humidity ranges was trend to approaching the mean value with the nitrogen application rate increased.3.Based on the characteristics of GA-BP artificial neural network and the weight value among Input Layer-Implicit Layer-Output Layer,we calculated the degree each factor affected by the soil respiration.From the calculation results,it shows that the litter level was the most important factor which influenced the soil respiration,while the other three factors showed a consistent contribution rate.