Based On Ensemble Kalman Filter Dynamic Prediction For Forest Resources

Forest resource is a kind of important renewable natural resources, but also state-ownedresource and an important part of the national wealth, development of forest resources inmaintaining human future and destiny, which provide a material basis for human sustainabledevelopment, but also for human sustainable development to provide environmental foundation.Therefore, on forest resources for scientific and dynamic prediction, timely grasp of the currentsituation of the forest resources and dynamic change, sustainable forestry development needs, butalso the important content of forest management plan, to make proper forest managementstrategy and forestry development strategy provides scientific basis, for the forestry planning,reasonable development to use silvan resource provides important scientific basis. Based on theensemble kalman filter and double ensemble square root filter technology on forest resourcedynamic prediction, the specific contents are as follows:1. This paper introduces the basic kalman filter, extended kalman filter and the ensemblekalman filter calculation process. Review of the ensemble kalman filter development course,analyzes the advantages and disadvantages of the ensemble kalman filter, and the ensemblekalman filter and four-dimensional variational in theory are compared.2. Based on the ensemble kalman filter technology to construct forest state space model, andusing EM algorithm on the model parameter estimation, finally with the residual lattice model,relational model and variance ratio of qualified qualified model of state space model has beentested. The results indicate that the model is of high precision, so it can be used to predict thedynamic of forest resources.3. Based on the kalman filtering technique and the ensemble kalman filter technique tostudy the dynamic prediction of forest area, with1950to1998national census data as samples offorest area, forest of countrywide forest area of dynamic prediction. kalman filtering of forestarea to predict the maximum error is23%, ensemble kalman filter prediction of forest area of themaximum error is6%. kalman filtering technique based on forest area prediction error fluctuationratio based on ensemble kalman filter technique to forecast the fluctuation of large forest area.Based on the ensemble kalman filter technology in forest area prediction more accurate results,the model is more flexible and the large scale data faster computing speed.4. Based on the study of double ensemble square root filter assimilation technique,combined with forest space state model of double ensemble square root of forest stumpagevolume and forest stock volume prediction based on kalman filter technology, collection of livestumpage forest volume prediction and the maximum error is8%, the minimum error0.The total forest stumpage volume error of2.7%,the total forest stock volume error of2%.The error isrelatively stable, less fluctuation.5. Ensemble kalman filter and double ensemble square root filter comparison of predicted,based on double ensemble square root filter technique for prediction of forest volume maximumerror is6%, based on the collection of kalman filter to predict the maximum error of8%. Doubleensemble square root filter technology forecasting error fluctuation. relatively stable, than theensemble kalman filter technology to predict the effect of good.