Thermal Infrared Multi-scale Observations And Response Mechanism Of Pests In Coniferous Forest

At present,remote sensing technology is an effective method to monitor diseases and insect pests and evaluate forest health in large-scale.Optical remote sensing based on spectral characteristics of ground stands to identify the health status is the main technical methods in major research.However,the monitoring accuracy of green attack stage is not satisfied the requirement,and the potential of thermal infrared remote sensing has not been explored.In this paper,the feasibility of monitoring by thermal infrared technology was analyzed based on the multi-scale observation results of pests in coniferous forest.The temperature change mechanism of damaged needles was discussed and analyzed utilized the observation data at leaf scale and SPA model,and the pest stress module was expanded to simulate the temperature of damaged leaf and soil components.In order to provide threshold temperature and operational analysis model for thermal infrared remote sensing in forest pest monitoring system,the pest stress module and three-dimensional heat radiation transfer model RAPID was coupled to simulate the directional brightness temperature of pest stressed stands.The research results mainly include the following aspects:1.Potential assessment of thermal infrared technology for monitoring pest stress.(1)At leaf scale,the temperature differences?T between the damaged needles and healthy needles reached the maximum from 14:00 to 15:00,and?T of the slightly,moderately,and severely damaged needles reached 0.6?,0.7?and 2.5?respectively.The physiological and biochemical factors such as CCL,WCL,P_n,g_s and T_rdecreased with the increase of the damaged degree.CCL and WCL were negatively correlated with NDVI and NDWI,while g_sand WCL were negatively correlated with?T.(2)At canopy scale,LAI was negatively correlated with canopy temperature and plot average temperature(R~2=0.8313,0.4143).When LAI<1,LAI had great interference on the relationship between temperature and SDR.The canopy temperature and plot temperature with LAI?1 had significant positive correlation with SDR(R~2=0.7371,0.5011).(3)At pixel scale,SDR was positively correlated with LST.The correlation difference between LST and SDR of Landsat 8 and GF-5 in the same region was small.The results showed that the accuracy of SDR inversion by NDVI and LST of GF-5 is higher than Landsat 8,and the accuracy of SDR inversion by NDWI and LST from different data sources is similar.In addition,the single band of pixel was hardly to reflect the SDR change with time.The spectral index constructed by linear and nonlinear combination of different bands could reflect the difference of vegetation characteristics under different SDRs.When SDR was small,vegetation indexes(SAVI?MSAVI?MSAVI2?NDVI)were negatively correlated with SDR,and water indexes((NDWI?MNDWI)were positively correlated when SDR was large.2.Temperature simulation and verification of different components under pest stressThrough the detailed analysis of SPA model,it was found that the greatest variable that affected the leaf simulated temperature in SPA model is the leaf net radiation_nR.On this basis,the pest stress module was developed to simulate the leaf and soil temperature.The results showed that the simulation accuracies of slightly damage and healthy conifers were higher,and the errors were smaller(RMSE=0.69k,1.12k).The simulation results were closer to the measured values at noon.The temperature difference between light and shadow soil was obvious due to solar radiation.There was no significant difference of soil component temperature under different damaged canopy.3.Simulation and verification of three-dimensional forest temperature distribution under pest stressed by coupling RAPID and SPA model.Taking the UAV thermal infrared image as reference,the coupled model could simulate the three-dimensional temperature of damaged pinus yunnanensis stands(R~2=0.8343,RMSE=0.7028K).The sensitivity analysis of LAI and SDR was carried out based on the coupling model.The results showed that the canopy temperature decreased with the increase of LAI and increased with SDR.When LAI<2,SDR had little effect on canopy temperature,and canopy temperature changes more rapidly with LAI.when LAI?4,scene temperature changes little with LAI.When LAI>10,canopy temperature tends to be stable.