Evaluation of Canopy Nitrogen in Dryland Wheat using ETM+ Satellite Images in South of West Azerbaijan Province

In cereal crops, nitrogen is the most important element for maintaining growth status and enhancing grain yield. Nitrogen is an important constituent of the chlorophyll molecule and the carbon-fixing enzyme ribulose-1, 5-bis-phosphate carboxylase/oxygenase. Therefore, providing enough nitrogen to achieve optimal yield is essential. Common chemical analyzes are used to determine the nutrient elements of plants using laboratory methods. Conventional laboratory techniques are expensive, laborious, and time-consuming. Determination of plant biochemical content by remote sensing could be used as an alternative method  which reduce the problems of laboratory analyses. Expensive and time-consuming direct determination of the nutritional status of the plant play an important role in the quantitative and qualitative yield of the product. However, exposure to rainfed wheat nutrient stresses (in particular, nitrogen) compared to irrigated wheat resulting in attempts to evaluate these features with acceptable accuracy without the direct measurement. In this regard, remote sensing data and satellite images are of the basic dryland management and optimal wheat production methods. As such, it collects massive information periodically from the surface of the planet, and it is easy to use this timely information to identify the stresses and apply appropriate agronomic methods in order to counteract them or reduce their negative impact on the production of this strategic product. Therefore, the goal of this study was to determine the nitrogen concentration of dryland wheat in the laboratory and its fitting with ETM+ images, evaluate the accuracy of remote sensing in determining the total nitrogen content of the plant and establish a regression relationship to estimate the amount of canopy nitrogen in the plant.

This research was undertaken in parts of the south of the West Azerbaijan Province in Iran. The sampling was done from 45 dryland wheat fields using a stratified random method in May 2016. The wheat canopy nitrogen was determined using the Kjeldahl method. Satellite images of the ETM+ were downloaded on the USGS website. Then the required pre-processing was performed on images to reduce systematic and non-systematic errors. Statistical analyses were performed by excel and SPSS. Descriptive statistics and correlations were obtained between reflectance data obtained from various satellite bands and nitrogen measured in the laboratory. Correlated variables among the reflectance data of different bands were analyzed by principal component to reduce repeat calculations. The regression relationship between the plant canopy nitrogen and the first principal component has been evaluated using the stepwise regression method. To draw the plant canopy nitrogen, map, the equation was obtained and the ETM+ image has been used for land uses. Finally, the map of canopy N distribution at the studied area was drawn.

The results showed that nitrogen content varied from 1.6% to 0.79%, with an average of 1.11%. The normality data was verified by the Shapiro Wilk test. The results of the Pearson correlation showed that the wheat canopy nitrogen has a high correlation with digital number values of all bands of satellite images except band 4, so that it has the highest and the least correlation with band 2 and band 4, respectively. The correlation between remote sensing data in different bands was also evaluated using bi-plot statistics, which results showed a high correlation between all bands except band 4 with the first one of the principal component (PC1). Therefore, only PC1 data has been used to study the regression relationships between wheat canopy nitrogen and remote sensing data. A regression equation  between wheat canopy nitrogen and ZPC1 (R2= 0.71) was developed. ZPC1 is obtained according to the following formula:  where ZPC1 is the standardized Z parameter, is the average of PC1 and the ????pc1 is the standard deviation of PC1. Finally, the map of canopy N distribution was drawn to the studied area. According to the results of this study, the application of remote sensing data such as Landsat ETM+ data is a very important variable for improving and managing the prediction of wheat canopy nitrogen.

Overall, the results indicated that the remote sensing data provide more accurate and timely information from the drylands of Iran to manage farm fertilization and prevent the decline in yields at critical points. However, proper management to avoid the fertilizer loss by precise and timely application of N-fertilizer is needed.