Spectral reflectance simulation and estimation of chlorophyll and water content of Pistacia mutica leaf based on PROSPECT4 model

Zagros forests, as the largest forest ecosystem in Iran, are mainly composed of oak and pistachio species. Although Pistacia mutica has the ability to adapt to adverse environmental conditions, for optimal establishment and growth, like other tree species, it needs environmental conditions appropriate with its ecological needs. Diameter and height growth in these trees are slow due to climatic conditions, so the study of the quantitative and qualitative condition of the forest by measuring and monitoring quantitative characteristics will not be accompanied by accurate results. In contrast, studying the biochemical and biophysical properties of the leaves and canopy of these trees can provide a more appropriate way for studying and monitoring them. Plant chlorophyll and moisture are important parameters in determining the physiological status, health condition, and stress status of trees. It is possible to estimate these parameters from remote sensing and proximity data using radiation transfer models that work according to the physics laws and how electromagnetic radiation interacts with trees. The PROSPECT4 model is one of the newest models proposed to estimate the amount of chlorophyll, water content, and leaf dry matter per unit area based on spectral reflectance measurements. In recent years, fires, pests and diseases, climate change, and drought have gradually affected the growth and quality of pistachio species. Since the first signs of stress in trees appear in their leaves, in this study, the need to investigate the quantitative and qualitative status of this species based on the biochemical parameters of its leaves through non-destructive methods of proximity was considered.

20 Pistacia trees were randomly selected in the Kood Siyah forest of Felard section of Chaharmahal va Bakhtiari province. The amount of chlorophyll parameters, equivalent water thickness and leaf dry matter were calculated in the laboratory. Spectral reflectance of leaf samples was measured by SVC HR-1024 spectrometer. Spectral data and values of leaf biophysical and biochemical parameters were entered in the ARTMO toolbox. Then, PROSPECT4 radiation transfer model was used to simulate spectral reflection and estimate water and leaf chlorophyll of Pistacia mutica. Combining the simulations with least squares regression, the performance of PROSPECT4 model in estimating chlorophyll content and leaf water of this species was evaluated.

In order to evaluate the model in estimating chlorophyll content and leaf water content, R2 and RMSE indices were used between the measured and estimated values. The results showed that the PROSPECT4 model in combination with the PLS model has good accuracy in estimating leaf water content (R2 = 0.73, RMSE = 0.0028) and leaf chlorophyll (R2 = 0.72, RMSE = 2.61). The results of paired T-test of spectral indices showed that ARI,, ARI2,, DWSI,,NDWI and p550 indices were not significantly different between the measured and simulated reflectance.

Based on the results of this study, the combination of radiation transfer model with regression methods such as PLS has great power in predicting tree parameters. Estimation of forest quality parameters in a vast area of Zagros forests using satellite data along with other radiation transfer models using several tree species and a range of parameters, as well as techniques such as ancillary information, multiple solutions, and other regression methods of simulating spectral reflectance can be performed.