*DOI: http://dx.doi.org/10.5772/intechopen.103898 An Overview of Soil Moisture and Salinity Sensors for Digital Agriculture Applications*

sensors are potentially designed to monitor vast regions; however, their spatial resolution depends on the microwave frequency, antenna size, and elevation. Most passive radiometers have a spatial resolution of 10 km, which is inapplicable for hydrological aims. Although microwave remote sensing drives many algorithms for calculating large-scale soil moisture, their low resolution is not appropriate for small scale [106]. Presently, the passive microwave retrieved resolution of soil moisture is about 25 km [107], and the low spatial resolution outputs, unreliable rainfall, and evaporationtranspiration data can make it challenging to estimate irrigation water demand [5]. Moreover, soil moisture data may not be available regularly. The spatial distribution of soil moisture is a prerequisite for agricultural and ecological management, while retrieving soil moisture data in heterogeneous landscapes is a significant challenge [108]. Heterogeneous landscapes generate irregularities in moisture measurements [109]. Consequently, merging surface reflectance data and auxiliary geospatial data can accurately estimate soil moisture, supporting precision agriculture strategies efficiently. **Table 1** summarizes some investigations that measured soil moisture using a combination of proximal and satellite data.


#### **Table 1.**

*Some studies on the merged application of ground-based and satellite sensors to estimate soil moisture.*

Remote sensing data can be applied to map surface soil salinity in broad regions [39], and the Landsat satellite has made it attainable to study soil salinity at different scales [118]. Wu et al. [119] reported that the overall accuracy of Landsat in soil salinity detection from 1973 to 2006 was approximately 90.2%. Combining proximal instruments with remote sensing systems is advantageous in precision evaluating soil salinity [120]. Bouaziz et al. [121] extracted 18 indicators from MODIS Terra data to improve salinity prediction patterns in northeastern Brazil and recognized a moderate correlation between EC and spectral indices. However, the limitations of using remote sensing data to map salt-affected areas include salt spatial distribution, temporal changes, and vegetation interference [122]. Moreover, it is challenging to estimate soil salinity through single-factor models [123]. Although remote sensing has numerous advantages over conventional proximal systems for mapping and predicting soil salinity [124], it is possible to determine the spatial variability of soil EC by local proximal sensor EM38 connected to GPS [125]. Casterad et al. [126] applied a combination of soil experiment data, proximal sensors, and satellites to investigate how soil salinity develops and distributes. Corwin [127] used proximal sensors and remote imaging to assess soil salinity at different scales; furthermore, Douaoui et al. [128] demonstrated that the regression-kriging approach combines remote sensing systems and ground network monitoring stations, thereby providing well-defined spatial and temporal monitoring of soil salinity. Eldeiry and Garcia [129] similarly reported that the modified kriging model presents the most reliable estimate of soil salinity by combining satellite and proximal data.

In a study by Fourati et al. [52], ordinary kriging with an average of 1.83 squares and a standard error of 0.018 had the most reliable performance for identifying and classifying saline soils. In an investigation by Fan et al. [130], the partial least squares regression model was applied to retrieve soil salinity from multispectral sensors, allowing salinity mapping with low cost and significant accuracy. Yahiaoui et al. [131] analyzed the topographic characteristics of the study area using Landsat 7 satellite data; accordingly, they created a multiple linear regression based on height and an adjusted soil salinity index that could predict soil salinity by 45%. Soil salinity modeling by satellite and proximal data in central Iraq revealed that models could reliably forecast salinity with 82.57% precision [119]. Therefore, modeling spatial soil salinity changes based on remote sensing data regression analysis is an economical, simple, and promising approach [132].

### **4. Wireless sensing and IoT monitoring**

The precision agriculture approach employs new technologies to optimize farming inputs and ameliorate agricultural systems [133]. As one of the newest Internetbased technologies to have joined the agricultural sector, IoT is a type of intelligent sensor with software based on a web connection, applied to proposed purposes on farms. It drives modern agriculture toward the automatization of manual operations [134, 135], and its architecture is shown in **Figure 8**. The Wi-Fi module forwards the soil parameter data assembled by the sensors to the controller and processor [136]. Growers can inspect soil moisture, temperature, and pH data on an Android mobile phone using IoT technology [137]. Automated irrigation can also minimize human mediation [138] as an incentive to save more water [139]. Yamin et al. [140] demonstrated that a digital soil test kit connected to the IoT system could be used to dynamically evaluate changes in soil elements. Moreover, IoT can help optimally control

*DOI: http://dx.doi.org/10.5772/intechopen.103898 An Overview of Soil Moisture and Salinity Sensors for Digital Agriculture Applications*

#### **Figure 8.**

*Deployment of hybrid data logger with Wi-Fi connectivity for IoT monitoring of soil moisture in berry fields. Source: SunBot.de.*

#### **Figure 9.**

*Wireless monitoring of soil moisture with solar-powered modular sensors.*

greenhouse conditions [141]. Shamshiri et al. [142] applied a systematic approach to automatically retrieving and processing greenhouse condition data in order to enhance tomato yield. Divyavani and Rao [143] could receive moisture sensor data using the Android mobile phone. Payero et al. [144] controlled soil moisture in a field through a mobile-based IoT system. The WSN system proposed by Shylaja and Veena [85] dispatched soil fertility circumstances to the mobile phone that are beneficial for fertilizer recommendation.

**Figure 9** demonstrates a solar-powered hybrid (Wi-Fi, LoRa, data logger) soil moisture and salinity sensors that were deployed in commercial berry fields in Germany. This device benefits from an onboard memory module for logging the measurements before transmitting the data via Wi-Fi and LoRa. It should be noted that due to the rising salinity trend caused by climate change, these devices are required for the precision monitoring of soil salinity in small and large scales [128]. Evaluating salinity-affected zones combats global climate change and prevents water resource loss [145]. Soil mapping is crucial for determining positional salinity levels and promoting appropriate management strategies for saline land restoration [146]. Therefore, combining remote sensing systems and EM38 sensors has provided an accurate soil salinity assessment approach, which is necessary to prevent further land salinization [76]. Future studies should concentrate on advancing remote sensing technologies for soil properties and the integration of salinity maps [147]. The measurement of soil moisture is critical in predicting drought and warning of natural disasters. Recently, many attempts have been made to address the development of soil moisture measurement facilities [148]. Launching advanced satellites promotes new innovative research approaches and encourages the development of new systematic empirical techniques for measuring soil moisture [149]. Non-cost-effectiveness plus inaccessibility to soil characteristics is one of the most significant constraints of precision agriculture [150]. Future soil moisture sensors should have high precision, low cost, and nondestructive features. Prospective research should also include the creation of specialized sensors for specific situations [33]. Using soil probes is critical for the most efficient and cost-effective use of water and chemical fertilizers; thus, numerous experiments on soil health indicators, such as water-holding capacity, salinity, temperature, pH, and soluble gas concentrations, are carried out [151]. However, high costs and the complex protection of sensors prevent the development of digital farming technologies, especially in rural regions [152].

For the purpose of downloading data from multiple sensors, a standalone software application shown in **Figure 10** was developed by Adaptive AgroTech to interface with the sensors' controllers via multiple serial COM ports as well as to execute commands and set custom configurations. The software also provides users with other features such as downloading log files of the sensor performance (i.e., battery and clock status, or historical parameters) or uploading the stored data to a cloud server. In addition, users can set labels to each node for simultaneously reading and writing log files from multiple devices and store the data on local memory cards. The Adaptive AgroTech Port Logger was developed in C# programming language environment and the Microsoft dot Net Core technology and can be operated on Microsoft Windows, Apple macOS, and Linux operating system. It should be noted that the MS-DotNet is a free open-source software for cross-platform development that supports various languages, such as C#, C++, and VB.NET. These features have provided a cost-effective and flexible solution for the future improvement of the Port logger. To have the best result and optimum performance, the software uses multithreading technology to execute parallel routines such as listening to multiports and executing more than one task at a time. Each thread defines a unique flow of control. As soon as the port logger engages in complicated and time-consuming parallel operations, it automatically sets different execution paths or threads, with each thread performing a particular task.

For the purpose of a visual comparison between air temperature, soil temperature, and soil surface moisture, sample data from the hybrid data logger shown in **Figure 9** that were collected every 10 minutes for 13 days in March 2021 are plotted in **Figure 11**. These plots validate the sensitivity of the sensor for the continuous monitoring

*DOI: http://dx.doi.org/10.5772/intechopen.103898 An Overview of Soil Moisture and Salinity Sensors for Digital Agriculture Applications*


#### **Figure 10.**

*Adaptive AgroTech Port Logger software for simultaneously downloading data from multiple sensor nodes under windows and Linux operating system.*

**Figure 11.**

*Plots of air temperature, soil temperature, and soil surface moisture during 13 days of experiment for performance evaluation of an adaptive AgroTech hybrid data logger.*

of agricultural field and for planning precision irrigation practices in arid areas. The measurements of the hybrid data logger can be used as the feedback data for a decision support system or controller that activates the irrigation pumps based on air and soil temperature, soil moisture, hours of the day, and other field parameters. It can be seen from the plots of **Figure 11** that during early morning hours, soil surface experiences more moisture (due to the morning dew) in the entire 13 days of the

experiments compared to the mid-day hours. It can also be seen that the hybrid datalogger did not miss a single measurement during the experiments, even when the air temperature was below the freezing point.
