Study On Water Deficit Of Walnut Forest Based On Infrared Thermography Information

As the most important economic tree species and woody oil tree species,walnut(Juglans regia L.)are widely planted in the whole world.China’s walnut production ranks first in the world.The ensuing irrigation water problem in the walnut plant orchard has become increasingly prominent.At present,the water management technology of the walnut plantation in the hilly area of North China is relatively extensive,and there are widespread problems such as unsuitable irrigation timing and excessive irrigation,which leads to waste of water resources.One of the keys to solving the problem is to timely and accurately diagnose the soil moisture status of the walnut trees and the degree of plant water deficit.The real-time and non-destructive measurement of canopy temperature can be realized by infrared thermal imaging technology.The study of plant water status based on canopy temperature is of great significance for exploring the effects of water on the physiological processes of plant growth and development.In addition,the constructed plant water deficit map is used to diagnose the water condition of the plant,combine the soil water content to achieve reasonable irrigation and optimize the water resources allocation of the orchard.The main conclusions are as follows:1.Uncertainty analysis of canopy temperature measurementIn order to explore and quantify the uncertainty caused by the thermal infrared imaging system for the measurement of walnut canopy temperature.Firstly,the sensitivity analysis was performed,and the degree of influence on the canopy temperature is from large to small as follows:leaf emissivity>ambient reflection temperature>air temperature>air relative humidity>distance.Secondly,two-way analysis of variance and the multiple comparison analysis were performed on the canopy temperature of three sample trees and four different directions.The results showed that there was a significant difference(p<0.05)between the south and north directions,but the differences between the different angles were not significant.Then,the temperature frequency histogram reflects that the thermal infrared image exhibits a bimodal distribution characteristic,in which the peak temperature of the canopy pixel is25.1°C.Thirdly,the analysis of the variance of the temperature inside and outside the canopy showed that there was a significant difference(p<0.01)between them,and the external temperature of the canopy reached the maximum at 13:00 in the afternoon.Uncertainty analysis of canopy temperature provides a theoretical basis for improving measurement accuracy.2.Accurate extraction of canopy temperatureIn order to accurately obtain the canopy temperature of walnut,this study elaborates and verifies the proposed optimal pixel extraction method.Firstly,in the entire region of interest,(ROI)regions,it was found that there is an inflection point when the gray value is 95 between the number of pixel points and the average temperature,and the inflection point is considered as the threshold value of the pixel point extraction from the inside canopy.Then,the coefficient of variation of the pixel points extracted by the adjacent gray value was calculated and combined with three kinds of data filtering algorithms to smoothly reflect change characteristics of the coefficient of variation.It is found that when the gray value is equal to108,the three data filtering algorithms intersect with the real values,which is considered to be the threshold for extracting the pixel points of the outside canopy.Above this threshold,the soil pixels are extracted,so the optimal pixel point for canopy temperature extraction is 12119.Finally,the verification results of the optimal pixel point extraction method in different spatial scales show that the coefficient of variation is only 1%,which has good stability.Comparing with the measurement results of the thermocouple,it is found that the temperature difference between them is not significant,indicating that the method has higher extraction accuracy.3.Establishment and application of soil water prediction modelIn order to scientifically manage the irrigation water use of walnut orchards in arid or semi-arid regions of northern China,the thermal infrared images of the sample trees and different water treatment plots were obtained by using the fixed thermal infrared camera A310f and the unmanned aerial vehicle thermal imaging system TC640,respectively.The results show that optimum time period for soil water prediction is considered from 13:00 to 14:00 in the afternoon,and there is a significant separation between the canopy-air temperature differences during this period.The canopy temperature is significantly higher than the air temperature and the variation range is 3°C to 5°C.The canopy-air temperature differences are negatively correlated with soil water content and positively correlated with solar radiation,where the contribution of soil water content is 75%.Therefore,using the canopy-air temperature differences to establish a soil water prediction model,R~2=0.64,p<0.001,RMSE=0.4,indicating that the model has a certain fitting accuracy.The established model was verified by actual measurement data,R~2=0.61,p<0.001,RMSE=0.3,indicating that the model has better prediction ability.Finally,the model is applied to the prediction of regional water conditions,which proves that it has a good application effect.4.Diagnosis of walnut water deficit by using crop water deficit index(CWSI)To assess the ability of both empirical CWSI(CWSIE)and analytical CWSI(CWSIA)methods to diagnose walnut water deficits and the feasibility of soil relative available water(RAW)mappings for guiding irrigation.The CWSIE and CWSIA mappings of the irrigated and non-irrigated areas were constructed,respectively.From the overall area,both methods can reflect the water deficit of walnuts.However,from the individual sample trees,there are obvious spatial variability characteristics.The CWSI is divided into four parts:1>CWSI≥0.7,under severe water stress;0.7>CWSI≥0.4,under moderate water stress;0.4>CWSI≥0.2,under mild water stress;0.2>CWSI≥0,under non-water stress.Comparing the RAW maps based on different models,when the soil water is sufficient,the RAW map based on CWSIE is more sensitive to water changes.When soil water is deficient,the RAW map based on soil water prediction model is closer to actual values.