Estimating soil moisture using vegetation cover indices and soil surface temperature in agricultural and saline fields of Qazvin plain

Abstract

Introduction

Soil moisture is a key indicator for defining and identifying agricultural drought. Estimating soil moisture is useful for identifying water scarcity conditions in the early stages and evolving drought events, which have implications for crop yield uncertainty, food security, agricultural insurance, policy-making, and crop planning. Soil moisture can be estimated through various field measurement methods, which offer a wide range of techniques. Point-scale measurements are the most accurate methods for measuring Soil Moisture Content (SMC) and can also be fully automated. However, installing and maintaining these instruments can be time-consuming and costly. Soil moisture is a key indicator for defining and identifying agricultural drought. Estimating soil moisture is useful for identifying water scarcity conditions in the early stages and evolving drought events, which have implications for crop yield uncertainty, food security, agricultural insurance, policy-making, and crop planning This is particularly relevant for dry and semi-arid regions worldwide. Agricultural drought acts as a catalyst, leading to social and political conflicts in developing countries.

The advantage of using remote sensing data is the ability to create a large archive of high-resolution data.High soil salinity has negative effects on soil structure, nutrient content, and plant growth, leading to reduced crop yields and increased desertification. Various factors such as inadequate irrigation, excessive use of fertilizers, and land-use changes can contribute to increased soil salinity levels. Climate variations also play a significant role in increasing salt content in the soil, particularly in areas with low water levels and decreased groundwater quality. Therefore, monitoring soil salinity levels is crucial for sustainable soil and agricultural management.



Remote sensing has proven to be a suitable method for monitoring salinity in large-scale and heterogeneous landscapes. Through the analysis of remote sensing data, maps and spatial models of salinity and moisture distribution can be generated, aiding land management and developing risk reduction strategies. Modern remote sensing technologies include satellite imagery and aerial photographs, which provide valuable information on vegetation cover, soil composition, and moisture.



Over the past few decades, research in remote sensing has made considerable progress, and various tools for data collection and analysis have been developed. Among these tools, multispectral imaging sensors receive information across different wavelengths, providing higher accuracy and resolution in images.



Soil moisture plays a crucial role in regulating runoff, vegetation production, and evapotranspiration, making it essential for identifying agricultural drought. Estimating soil moisture can be used in the early detection of water scarcity and drought events. One common approach for estimating soil moisture is the use of remote sensing data. Satellite imagery and aerial photographs can provide useful information about soil color, texture, and vegetation cover, aiding in the estimation of soil moisture.







Materials and Methods

In this research, Landsat 8 and Sentinel-2 satellite imagery from February 2023 were used to estimate soil moisture. The Google Earth Engine platform was utilized for processing and calculations. Various vegetation indices and the Land Surface Temperature (LST) index were analyzed to assess their relationship with soil moisture. Ground data was collected using the HH2 Moisture Meter device for 23 soil moisture samples.

Results and Discussion

The results showed that there was a higher correlation between vegetation indices derived from the SENTINEL-2 sensor compared to the LANDSAT-8 sensor. Indices such as NDVI, SAVI, and NDTI had a high correlation with soil moisture content, with NDTI showing the highest correlation of 0.84. Based on the indices with the highest correlation, a regression model was developed to estimate soil moisture content. The results indicated that the regression model using the Land Surface Temperature (LST) and NDTI indices from the LANDSAT-8 sensor had the highest accuracy, with a coefficient of determination (R-squared) of 0.81 and a bias of 0.27.

Conclusion

The results of this study demonstrate that the use of remote sensing data, particularly LANDSAT-8 and SENTINEL-2 satellite imagery, can be an effective tool for estimating and monitoring soil moisture and soil salinity in agricultural and saline areas of Qazvin plain. Furthermore, continuous utilization of remote sensing data at short time intervals can serve as a useful tool for accurate and continuous soil moisture monitoring.



In this study, the correlation matrix between the calculated indices from satellite images and soil moisture was examined. The results showed that all indices, including SAVI, NDVI, NDTI, NDMI, and SMSWIR, have high correlations with soil moisture. In particular, SAVI and NDVI indices in LANDSAT-8 images had the highest correlation with negative values of 0.84 and 0.71, respectively. Moreover, the correlation analysis of land surface temperature with vegetation indices demonstrated high correlations for all indices.



In addition to correlation analysis, regression models were developed to estimate soil moisture using two indices, NDTI and LST. These models utilized influential factors on soil moisture and were capable of more accurate estimation of soil moisture using LANDSAT-8 images. Therefore, these models were introduced as regression models with high accuracy within the study area.



In general, the results of this section of the study indicate that the use of remote sensing data, particularly LANDSAT-8 and SENTINEL-2 satellite images, is beneficial for soil moisture estimation and salinity monitoring in agricultural and saline areas of Qazvin plain. Furthermore, land surface temperature serves as a useful indicator for predicting soil moisture and analyzing its temporal changes. The use of regression models based on calculated indices from LANDSAT-8 and SENTINEL-2 satellite images can lead to high accuracy in estimating soil moisture and soil salinity in the study areas.



By utilizing remote sensing data, soil moisture can be continuously monitored, and its changes over time can be observed. This information can assist farmers and water resource managers in determining the optimal timing for irrigation and efficient water resource management.



Overall, the use of remote sensing data in estimating and monitoring soil moisture and soil salinity in agricultural and saline areas of Qazvin plain can significantly improve water resource management and agriculture for this region.