Advanced Autonomous System for Monitoring Soil Parameters

Abstract

Context: This research investigates the advantages of real-time monitoring of soil quality for various land management practices. It also highlights the significance of spatio-temporal soil modeling and mapping in providing a clear and visual understanding of how aridity changes over time and across different locations. Aims: This paper aims to provide a comprehensive guide to the key processes required for the development of a laboratory-based soil quality monitoring system. Methods: The applied methodologies involved the processes of sensor deployment, data acquisition infrastructure establishment, and sensor calibration. These procedures culminated in the development of a soil quality assessment model that was subsequently subjected to two months of laboratory testing using three distinct soil types. The analysis yielded a strong positive linear correlation between the measured and predicted soil quality values. Key Results: As expected, the assimilation of prior soil quality estimates within the modeling framework demonstrated a significant enhancement in the accuracy of real-time soil quality estimations. Conclusions: This research promotes the importance of iterative improvements of the soil quality monitoring system. The need for a long-term perspective and a plan for maintenance and continuous improvement of such systems in the ecosystem is important to improve the ease of making predictions to avoid soil aridization. The results of this research will be useful for researchers and practitioners involved in the design and implementation of soil monitoring systems.

1. Introduction

Accurate monitoring of key soil parameters, including temperature, moisture, pH, electrical conductivity (EC), and nutrient levels (NPK), is crucial for predicting soil health and developing a comprehensive real-time understanding of soil conditions across any given area. This necessitates the adoption of advanced technological systems.

This shift towards advanced technology is a concerted effort to simultaneously conserve natural resources and enhance profitability by streamlining production processes and minimizing resource consumption. As regions grapple with dwindling water supplies and the increasing threat of drought, technology assumes a critical role in enabling accurate predictions and proactive mitigation strategies to avert water scarcity and its associated consequences.

Soil nutrients are critical for agricultural productivity, directly influencing plant growth and yield. Concurrently, soil moisture significantly impacts soil erosion processes as well as water and sediment transport within the soil profile [1]. These macronutrients, including nitrogen (N), phosphorus (P), and potassium (K), are essential for various biochemical processes within plants. Notably, nitrogen and phosphorus are crucial for plant growth and play a pivotal role in the soil–plant system [2]. While nitrogen and phosphorus are primarily found within the soil, their concentrations exhibit significant variability influenced by factors such as soil type, geographic location, and prevailing environmental conditions [3]. Soil factors such as temperature, moisture content, water table depth, and electrical conductivity are critical determinants of crop quality and yield. Precision agriculture can be defined as the strategic application of modern information technologies to acquire, process, analyze, and optimize the management of agricultural resources [4]. This directly involves the collection and analysis of site-specific data to enable informed and individualized management decisions that mitigate soil aridization and, crucially, prevent desertification.

2. Literature Review

This research commenced in 2023 with the initiation of the PN 23 43 05 02 “IoT-SOL” project at INCDMTM-Bucharest, spanning a four-year duration. The development of this manuscript was facilitated by the invaluable contributions of colleagues from the Departments of Environmental Engineering and Renewable Energy Systems, Intelligent Thermo-technical Measurements, and Complex Automation and Control Systems. The author declares no conflicts of interest regarding this research, which focuses on “advanced autonomous system for monitoring soil parameters.” This work addresses the development of technological solutions by applying principles from natural sciences and carefully selecting appropriate materials. The reviewed literature, as cited, explores different aspects compared to the present study, such as:

Padmavathi et al. [3] present an IoT framework for agriculture, focusing on pest detection, automated irrigation, and alert systems. This framework necessitates the deployment of various sensors and IoT devices. In contrast, our research specifically focuses on measuring key soil quality parameters—temperature, moisture, electrical conductivity (EC), and nutrient levels (NPK)—at varying soil depths.

Rudi et al. [5] present the design and implementation of a portable IoT-based system for real-time monitoring of soil nutrients (nitrogen, phosphorus, and potassium) using field-based sensors. This system aims to enhance precision farming by enabling on-site soil analysis and providing tailored fertilization and management recommendations. In contrast, this study focuses on analyzing soil nutrient levels at various depths (up to 80 cm) to generate data for predicting and mitigating soil aridization.

We investigate the key processes involved in developing an operational on-farm system for high-resolution daily soil moisture mapping, capturing temporal, spatial, and vertical variations. The paper emphasizes the importance of iterative improvements, particularly in sensor recalibration and spatio-temporal modeling. In contrast, this study focuses on analyzing soil properties at various depths (up to 80 cm) to generate data for predicting and mitigating the risk of soil aridization.

Currently, this research is in the laboratory phase, where various data acquisition, transmission, storage, and display technologies are being evaluated. The objective is to identify the most suitable combination of technologies for the specific conditions of in-situ equipment installation. The acquired data will be stored and visualized on an Ecosystem IoT platform developed specifically for this project.

The proposed multisensory soil monitoring system is designed for easy integration into various substrates and soils through a compact probe design, enabling rapid assessments and optimization of cultivation practices. Moreover, this system facilitates the creation of geo-temporal maps of soil conditions across different regions, providing crucial data for informed decision making to mitigate soil aridization. The proposed autonomous, multisensory system will enhance the precision of digital soil mapping by providing a sustainable and renewable means of geographically characterizing the aridization status of various regions, ultimately supporting sustainable soil management practices.

The primary objective of this work is to develop a laboratory system capable of simultaneously acquiring data from three distinct soil types to enable comprehensive monitoring of soil characteristics and mitigate the risk of soil aridization.

The advantages of IoT-based soil monitoring, as illustrated in Figure 1, extend beyond immediate crop management. By collecting and analyzing historical soil data, it becomes possible to identify patterns and trends, leading to a deeper understanding of soil behavior and properties. This knowledge can then be leveraged to develop long-term soil management plans that prioritize soil health, nutrient cycling, and water conservation.

3. Materials and Methods

3.1. Overview

Soil quality assessment plays a crucial role in maintaining soil health, particularly in mitigating the risks of aridization and desertification. Prior to establishing any crop, thorough soil testing is essential. This practice enables farmers to determine the suitability of the soil for specific crops and to develop appropriate fertilization plans. Ideally, soil testing should be conducted at least once every four years to ensure ongoing soil health management [6].

One of the primary functions of soil is to produce phytomass, the raw material for food, clothing, fuel, and other essential products. This function is crucial for human sustenance and economic development. Enhancing soil health through practices like reducing energy consumption and adopting electric tractors is essential for optimizing this vital function [7].

For optimal plant growth, soil provides a diverse array of chemical elements essential for vegetation development and crop production. Among these elements, fourteen are recognized as essential nutrients for plants. Based on the quantities required by plants and their specific physiological and biochemical functions, these nutrients are categorized into two groups: macronutrients and micronutrients [6]. Macronutrients are further categorized into primary macronutrients (nitrogen (N), phosphorus (P), and potassium (K)) and secondary macronutrients (sulfur (S), calcium (Ca), and magnesium (Mg)) [8].

Nitrogen (N) is a significant component for plant growth, as it stimulates root growth and plant development as well as the absorption of other nutrients. A crop with a nitrogen deficiency will have reduced vegetative growth, discolored leaves, and poor yields, strictly correlated with the strength (or lack thereof) of the plants in the crop that have a nitrogen deficiency.

Phosphorus (P) is primarily a valuable source of energy; it is a valuable constituent of ATP (adenosine triphosphate), the molecule that provides the energy needed for a fairly large number of metabolisms. It is also a constituent of cell membranes and DNA molecules. This central role makes phosphorus an essential nutrient for plant structure and growth.

Potassium (K) plays an essential role in plants, participating in essential processes such as photosynthesis, cellular respiration, opening and closing of stomata, water absorption, protein metabolism, nutrient absorption and utilization, cell wall strengthening, and enzyme activation. As a result, this mineral is an essential nutrient for proper plant growth and development and must be controlled in any type of crop to ensure the most efficient agricultural production.

The group of micronutrients includes: Iron (Fe), Manganese (Mn), Cobalt (Co), Copper (Cu), Zinc (Zn), Boron (B), Molybdenum (Mo), and Chlorine (Cl) [9,10]. To accurately determines the macro- and micronutrient requirements of soils, it is essential to assess their nutrient content. This assessment is typically conducted as part of agrochemical mapping activities. The process generally involves three distinct phases:

(1)

Field sampling, including the identification and sampling of target areas;

(2)

Laboratory analysis, which involves time-consuming and often costly procedures;

(3)

Data processing and decision making, which occur in an office setting.

Soil analysis is crucial for:

−

Determining soil type: Identifying the specific soil type is fundamental for understanding its characteristics and limitations.

−

Selecting appropriate crops: Determining the suitability of a particular soil for specific crops ensures optimal growth and yield.

−

Optimizing fertilization programs: Soil analysis helps determine the precise amount of fertilizer required for each crop, maximizing nutrient utilization and minimizing nutrient losses.

−

Assessing and mitigating risks: Soil analysis helps assess the risk of nutrient deficiencies or excesses as well as the potential for soil degradation, including aridization and desertification.

−

Determining key soil properties: Soil analysis provides vital information on soil pH and other essential indicators, such as organic matter content, which are crucial for understanding overall soil health.

Soil sampling should be conducted under dry conditions to avoid inaccuracies in nitrogen level estimations. Ideal soil texture for sampling is friable, allowing it to be easily crushed in the hand without sticking. Before sampling, it is crucial to remove surface organic debris to a depth of 2–3 cm.

By deploying smart sensors at various depths (e.g., 20 cm, 40 cm, 60 cm, 80 cm), we can continuously monitor soil conditions and weather parameters in real-time, eliminating the need for frequent manual sampling. This approach offers significant advantages in terms of efficiency, cost-effectiveness, and data accuracy.

Following the framework outlined by Brendan et al. [11] and considering the importance of technological innovation in system design [12], a soil monitoring system can be established through the following key steps:

• Define measurement parameters: Select the specific soil characteristics to be monitored based on the project’s objectives (e.g., temperature, pH, moisture, macronutrients like NPK).
• Choose sensor technology: Select appropriate smart sensors for the chosen parameters considering factors like accuracy, precision, cost, power consumption, and data logging capabilities.
• Implement data acquisition and transmission: Develop a reliable system for collecting sensor data at regular intervals. This might involve wireless transmission protocols for efficient data transfer.
• Establish data storage and processing: Set up a system for securely storing and processing the collected data. This might involve cloud-based storage solutions and data analysis tools.
• Sensor deployment and calibration:
  • Install smart sensors at various depths (e.g., 20 cm increments up to the desired depth) following proper installation procedures.
  • Integrate the system to manage sensor calibration and recalibration processes, ensuring accurate and reliable data acquisition.
• Monitor sensor performance:
  • Regularly monitor the health and functionality of the deployed sensors in the field.
  • Monitor and analyze the collected data to identify potential issues and ensure data integrity.

The current research primarily focuses on the initial stages of establishing a soil monitoring system, encompassing steps 1–4 outlined previously. These steps involve defining measurement parameters, selecting appropriate sensors, developing data acquisition and transmission systems, and establishing data storage and processing infrastructure.

Steps 5–6, which encompass sensor deployment, calibration, and ongoing monitoring, represent the practical implementation phase of the soil monitoring system. These aspects will be thoroughly investigated and analyzed as the research project progresses, providing valuable insights into the long-term operation and maintenance of the system.

LoRaWAN networks are known for their long-range communication capabilities, with outdoor coverage often exceeding 15 km in ideal conditions (line of sight, minimal interference) [13]. The soil sensors can use a battery with a solar panel, meaning they can be deployed over several hundred acres, providing a good indication of soil quality monitoring, or used for smart irrigation applications. Despite the benefits and extended coverage of LoRaWAN, this does not solve the problem of relaying data to a cloud platform or central system [14].

Soil quality is determined by using a soil sensor system that involves the integration of smart sensors to monitor a variety of characteristics, such as:

−

Soil temperature;

−

Soil moisture;

−

Soil conductivity;

−

Oxygen levels in the soil;

−

Soil water potential;

−

NPK sensors.

Buried electrode probes offer a reliable method for monitoring soil moisture content [15]. Soil moisture content is a cornerstone of hydrology, soil science, and agriculture. It significantly influences soil chemistry, plant growth, and groundwater recharge [16]:

−

Adequate soil moisture is essential for plant growth. It allows plants to absorb nutrients and water, which are crucial for photosynthesis and overall health.

−

Soil moisture content affects the movement and availability of nutrients within the soil. It influences chemical reactions and the leaching of nutrients

−

Soil acts as a natural filter, allowing water to percolate into the ground and replenish groundwater reserves. Soil moisture content directly impacts this process.

By understanding and monitoring soil moisture content, we can optimize agricultural practices, manage water resources effectively, and protect our environment.

3.2. Method

Smart soil monitoring sensors offer a comprehensive approach to soil analysis. These advanced devices can simultaneously measure multiple critical parameters, including temperature, moisture, nitrogen (N), phosphorus (P), potassium (K), volumetric water content, water potential, pH, and soil oxygen levels. The collected data are wirelessly transmitted to a central hub (or cloud platform) for real-time visualization and in-depth trend analysis. This enables farmers and researchers to accurately assess the degree of soil aridity, as illustrated in Figure 2, and make informed decisions regarding irrigation, fertilization, and other crucial agricultural practices.

The final research on soil quality monitoring will provide a comprehensive solution for collecting soil data and transmitting it through the local network to a computer for processing. Sampling depths will be determined according to current standards, considering the specific crops present.

Leveraging these data allows for optimizing farming operations, identifying trends, and making subtle adjustments to maximize both yield and soil quality. This integration of smart sensors into agriculture is known as Smart Agriculture (or Smart Farming) [17]. Within Smart Agriculture, the Internet of Things (IoT) plays a central role as a precision component by enabling real-time signal reception, transmission, and analysis to support informed decision making.

This research focuses on simultaneously receiving data from sensors and processing them to validate and calibrate the sensors using known soil samples.

The tests were carried out on three types of known soils as is presented in

Table 1. Technical parameters for the soil.

Parameters Soils to Analyze the Results	M1 *	M2 **	M3 **
pH (in water)	6.5	7.5	6.77
Electric conductivity (EC)	500 µS/m	340 µS/m	310 µS/m
Dry apparent density	250 kg/m3	133 mg/kg	0
N	99 mg/kg	63 mg/kg	237 mg/kg
P	242 mg/kg	260 mg/kg	97 mg/kg
K	130 mg/kg	0	393 mg/kg

* Producer: ORVITAR, Italy. ** Producer: Green Project SRL, Iasi, Romania.

:

−

M1-Floralis soil: A commercial potting mix designed for flowering plants, typically containing a blend of peat moss, perlite, and other organic matter.

−

M2-Soil for aromatic plants: A specialized soil blend formulated to support the growth of herbs and other aromatic plants, often enriched with nutrients and well-draining components.

−

M3-Planting soil: A general-purpose potting mix suitable for a wide range of plants, commonly containing a mix of peat moss, compost, and other organic and inorganic materials.

To conduct laboratory research, a portion of the architecture was implemented to acquire and display sensor data, as reproduced in Figure 3.

3.3. Interface

Analog interfaces are employed for sensors that generate analog outputs, including temperature, pressure, electrochemical sensors, light and IR detectors, thermocouples, touch controllers, and others [18]. To interface these sensors with microcontrollers, an analog-to-digital converter (ADC) is necessary to convert the analog signal into a digital format for easier processing by microcontrollers and other digital devices [19].

The RS485 interface is a differential serial communication standard designed for long-distance data transmission. It is widely used in industrial applications where numerous devices require interconnection within a system. A broad range of sensors, including those for ground or air applications, utilize the Modbus protocol over the RS485 interface.

RS485 is a communication standard known for its robustness, especially in noisy environments. It achieves this by transmitting data differentially, meaning the signal is encoded in the voltage difference between two wires [20]. This differential signaling helps to cancel out common-mode noise, which is noise that affects both wires equally.

This standard was designed with the aim of expanding and improving the communication possibilities on a serial line. Among the improvements considered, the following can be mentioned:

−

Increasing the maximum communication distance (approx. 1 km);

−

Increasing immunity to noise;

−

The possibility of multipoint communication (network communication);

−

The use of a cheap communication medium.

Characteristics of the RS485 bus:

−

A section of twisted bifilar cable is used, which has terminators (resistant) at both ends; both lines are used for transmission, there being no ground wire.

−

Coding of binary data is done by positive and negative differential voltages measured on the two lines of the section; the minimum differential voltage considered valid is +/−200 mV.

−

The emission circuits connected on the same section have tri-state outputs to allow several types of equipment to access the same communication section—access is multiplexed in time.

The RS485 standard allows the connection of up to 32 terminals (emission + reception). The RS485 standard does not specify the structure of the transmitted data, data flow control mechanisms, or error detection mechanisms. These can be taken from other serial communication standards (for example, RS232 or SDLC/HDLC) or can be defined by the user. This protocol is used as support for a series of protocols for industrial networks (for example, CAN, ModBUS, Profibus).

3.4. Smart Sensors

In this research, the smart sensors used (as illustrated in Figure 4) are sensors purchased from the ComWinTop company with the following technical parameters according to

Table 2. Technical parameters of smart sensor.

Characteristics	Accuracy
Temperature accuracy	±0.5 °C
Moisture accuracy	±3%
Conductivity accuracy	±10%
Accuracy of NPK	±2%
Salt accuracy	±10%
Measurement frequency range	50 Hz to 3 kHz
Nominal resistance	1000 Ω at 0 °C
Measuring current	0.3 mA
Connection	Pt/Ni wires, Ø 0.2 mm

[21].

The smart sensors utilized in this research were tested and calibrated at the Pressure Calibration Laboratory, accredited by RENAR within INCDMTM-Bucuresti. These sensors are a conductive type, exhibiting characteristics such as fast response time, resistance to chemical agents, excellent long-term stability, and an integrated RTD for temperature measurement and compensation.

3.5. Communication

Communication between the MASTER mode and the slave modules, responsible for data acquisition, is facilitated by the Modbus RTU protocol. This protocol defines a standardized message structure that controllers can readily recognize and utilize, regardless of the specific network type they are communicating with.

A communication protocol dictates how a controller requests access to another device, responds to requests from other devices, and handles error detection and reporting. It establishes a standardized format for message structure and content. The electronic data measurement and transmission equipment operates in two modes, remote and local, as illustrated in Figure 5.

Upon power up, the microcontroller firmware determines the operational mode: REMOTE or LOCAL. The default mode is LOCAL. In LOCAL mode, data acquisition is initiated by issuing a command from a PC terminal.

Preliminary sensor operation tests are conducted by directly connecting the sensors to a PC via an RS485-USB interface:

(1)

The MASTER module is connected to a PC using an RS232-USB interface adapter.

(2)

The operator enters ASCII commands in the Terminal.exe program. These commands are sent to the MASTER module. Alternatively, a macro-command can be created within the Terminal.exe v1.93b - 20141030beta - by Br@y++ application and executed by pressing a software button.

(3)

If the sent command is valid, the master module executes the program code associated with the command, as illustrated in Figure 5.

The flow chart, as illustrated in Figure 5, illustrates the workflow of a device interacting with other systems using the Modbus protocol. It operates in two modes: “Local” for autonomous data collection and processing and “Remote” for external control and data retrieval.

4. Calibration

Conductivity is a measure of the concentration of ions in an aqueous solution, such as moisture. It can be used to assess the quality and quantity of nutrients, salts, or impurities present in the solution [22,23]. The integrated temperature sensor enables both temperature measurement and compensation (if necessary). It can be strategically placed within the measurement area to enhance the accuracy of the acquired data [24].

Electrical conductivity in an aqueous solution is influenced by two key temperature-dependent factors [25]:

−

Ion concentration: As the temperature increases, the kinetic energy of molecules and ions increases, leading to increased molecular motion and potentially greater dissociation of salts into ions. This generally results in a higher concentration of ions in the solution.

−

Ion mobility: Higher temperatures typically increase the mobility of ions. As the temperature rises, the viscosity of the solution usually decreases, allowing ions to move more freely and conduct electricity more effectively.

When an electric voltage is applied between two electrodes immersed in an electrolyte (a solution containing ions), an electric current flows [26]. The voltage drop across an electrolyte is inversely related to its conductivity, which is significantly influenced by temperature. As the temperature increases, the conductivity of the electrolyte generally rises [27].

During the research, sensor testing and calibration software was obtained from the ComWinTop Store manufacturer. The calibration process involved determining the numerical coefficients A and B through curve fitting, specifically using a right-sided interpolation method.

Y = AX + B

(1)

where:

X is the raw measured value, and Y is the final measured value that is applied to the NPK sensors.

Sensor calibration was performed exclusively on the M1 soil type, as illustrated in

Table 3. Calibration values—I.

I	T [°C]	M [%rH]	pH	N [mg/kg]	P [mg/kg]	K [mg/kg]	CE [uS/cm]	TDS
	21.0	18.0	6.4	38.2	42.1	13.0	201.0	195.2
	21.1	18.0	6.5	38.4	42.3	13.3	201.3	195.6
	21.3	18.0	6.4	38.3	42.2	13.4	201.0	195.6
	21.1	18.0	6.5	38.1	42.1	13.2	200.0	195.3

T—soil temperature; pH—degree of acidity of the soil; M—degree of soil moisture; N—nitrogen; P—potassium; K—phosphorus; CE—the electrical conductivity level of water in soil; TDS—total dissolved solids in soil.

. The calibration procedure consisted of two steps:

• Initial calibration: Calibration was conducted using a solution of 200 mL distilled water with 2 g of NPK fertilizer.
• Soil calibration: Subsequent measurements were taken on 1 kg of M1 soil moistened with 100 mL of water

5. Results

This research involved conducting tests and measurements on all three soil types using the calibrated smart sensors.

Each sensor generates daily readings at eight-hour intervals. The working architecture, illustrated in Figure 5, processes these readings in conjunction with the initial calibration data. By applying a manufacturer-provided equation, the system estimates soil parameters, as outlined in

Table 4. Values measured on soils types.

	T [°C]	M [%rH]	pH	N [mg/kg]	P [mg/kg]	K [mg/kg]	CE [uS/cm]	TDS
M1	19.9	18.2	6.5	101.0	280.0	275.0	369.0	335.0
	19.8	18.4	6.4	96.0	268.0	262.0	354.0	322.0
	19.9	18.4	6.4	93.0	263.0	257.0	348.0	317.0
	19.8	18.3	6.4	98.0	274.0	368.0	361.0	329.0
M2	19.0	18.7	5.5	238.0	232.0	321.0	292.0	584.0
	19.0	19.6	5.5	238.0	232.0	319.0	290.0	581.5
	19.0	19.9	5.5	236.0	230.0	318.0	289.0	579.0
	18.9	18.1	5.4	240.0	234.0	322.0	293.0	587.0
M3	21.3	21.7	6.7	230.0	98.0	390.0	310.0	423.0
	21.4	21.6	6.8	231.0	97.0	393.0	308.0	420.0
	21.4	19.7	6.8	230.0	96.0	391.0	309.0	420.0
	21.4	20.5	6.8	229.0	97.0	390.0	310.0	421.0

.

The soil moisture sensor primarily responds to the clay content within the soil, which means that we expect similar values at various depths. Importantly, the calibration process does not significantly alter the observed pattern of soil moisture variation with depth. Recalibration does not affect the data model, resulting in soil wetting varying with the measurement depth in the soil profile in response to the contribution of meteorological conditions (amount of rain and high temperatures).

In this research, the correlation between the values obtained for the three soil types was determined using Pearson’s correlation coefficient. The following generic formula was employed to calculate the correlation coefficient between two variables, x and y:

$${\mathrm{r}}_{\mathrm{x}\mathrm{y}}={\displaystyle \frac{\sum _{\mathrm{i}}\left({\mathrm{x}}_{\mathrm{i}}-\overline{\mathrm{x}}\right)·\left({\mathrm{y}}_{\mathrm{i}}-\overline{\mathrm{y}}\right)}{\sqrt{\sum _{\mathrm{i}}{\left({\mathrm{x}}_{\mathrm{i}}-\overline{\mathrm{x}}\right)}^2·\sum _{\mathrm{i}}{\left({\mathrm{y}}_{\mathrm{i}}-\overline{\mathrm{y}}\right)}^2}}}$$

(2)

• n—sample size
• x—individual values of the variable x
• y—individual values of the variable y
• $\overline{\mathrm{x}}$—arithmetic mean of all x values
• $\overline{\mathrm{y}}$—arithmetic mean of all y values

Applying Colton’s rules (stated in 1974):

• *—A correlation coefficient from 0.25 to 0.50 (or from −0.25 to −0.50) means a weak correlation (acceptable degree of association).
• **—A correlation coefficient from 0.50 to 0.75 (or from −0.50 to −0.75) means a moderate to good correlation.
• ***—A correlation coefficient greater than 0.75 (or less than −0.75) means a strong correlation (very good degree of association).

Limits of the Pearson coefficient:

$$-1<r_{xy}<+1$$

(3)

Thus, as the value of the Pearson correlation coefficient approaches 1 (in absolute value), the “intensity” of the linear relationship between the two variables will be higher.

The experimental results demonstrated a strong correlation between NPK values and soil moisture. This relationship is evident in Figure 6, which depicts a clear positive linear correlation between moisture levels and key soil nutrients. Notably, soil type M3 exhibited a particularly strong positive linear correlation between these parameters.

The correlation values for M1, M2, and M3 soils are presented in

Table 5. Correlation of values for soil M1.

	T [°C]	M [%rH]	pH	N [mg/kg]	P [mg/kg]	K [mg/kg]	CE [uS/cm]	TDS
T [°C]	1
M [%rH]	−0.938 ***	1
pH	0.256 **	−0.429 **	1
N [mg/kg]	−0.989 ***	0.916 ***	−0.222 *	1
P [mg/kg]	−0.990 ***	0.927 ***	−0.238 *	0.999 ***	1
K [mg/kg]	−0.972 ***	0.893 ***	−0.279 **	0.977 ***	0.977 ***	1
CE [uS/cm]	−0.988 ***	0.915 ***	−0.000	0.999 ***	0.999 ***	0.978 ***	1
TDS	−0.988 ***	0.914 ***	−0.221 *	0.999 ***	0.999 ***	0.978 ***	0.999 ***	1

*—A weak correlation; **—A good correlation; ***—A strong correlation.

,

Table 6. Correlation of values for soil M2.

	T [°C]	M [%rH]	pH	N [mg/kg]	P [mg/kg]	K [mg/kg]	CE [uS/cm]	TDS
T [°C]	1
M [%rH]	−0.711 **	1
pH	0.993 ***	−0.691 **	1
N [mg/kg]	−0.997 ***	0.722 **	−0.996 ***	1
P [mg/kg]	−0.997 ***	0.722 **	−0.996 ***	0.999 ***	1
K [mg/kg]	−0.997 ***	0.723 **	−0.996 ***	0.999 ***	0.999 ***	1
CE [uS/cm]	−0.997 ***	0.711 **	−0.996 ***	0.999 ***	0.999 ***	0.999 ***	1
TDS	−0.997 ***	0.721 **	−0.996 ***	0.999 ***	0.999 ***	0.999 ***	0.999 ***	1

**—A good correlation; ***—A strong correlation.

and

Table 7. Correlation of values for soil M3.

	T [°C]	M [%rH]	pH	N [mg/kg]	P [mg/kg]	K [mg/kg]	CE [uS/cm]	TDS
T [°C]	1
M [%rH]	0.728 **	1
pH	0.813 ***	0.851 ***	1
N [mg/kg]	0.833 ***	0.927 ***	0.961 ***	1
P [mg/kg]	0.832 ***	0.932 ***	0.958 ***	0.999 ***	1
K [mg/kg]	0.834 ***	0.927 ***	0.961 ***	0.999 ***	0.999 ***	1
CE [uS/cm]	0.832 ***	0.926 ***	0.959 ***	0.999 ***	0.999 ***	0.999 ***	1
TDS	0.832 ***	0.928 ***	0.959 ***	0.999 ***	0.999 ***	0.999 ***	0.999 ***	1

**—A good correlation; ***—A strong correlation.

.

There is a general trend for macronutrients (N, P, K) to increase in concentration with increasing soil moisture, as is illustrated in Figure 6. In other words, the wetter the soil, the higher the nutrient concentration tends to be.

Strength of correlation:

N–M: The correlation between nitrogen and moisture appears to be moderate, with a linear increase in nitrogen concentration with increasing moisture.

P–M: The correlation between phosphorus and moisture is stronger than that between nitrogen and moisture, with a more pronounced increase in phosphorus concentration with increasing moisture.

K–M: The correlation between potassium and moisture is the strongest of the three, with an almost linear and consistent increase in potassium concentration with increasing moisture.

Data dispersion: The data for each nutrient show some dispersion around the trend line, indicating that factors other than moisture may influence the nutrient concentration in the soil.

Based on this graph, we can conclude that there is a positive relationship between M3 soil moisture content and NPK nutrient concentration. However, the strength of this relationship varies by nutrient, being strongest for potassium and weakest for nitrogen.

6. Discussion

This research has demonstrated the technology and processes required to enhance our understanding of soil quality through the implementation of a soil quality measurement network at various depths. Real-time monitoring of soil characteristics offers significant advantages in land management. By creating spatio-temporal maps based on data collected from a network of measurement points across different geographical areas, we can gain a clear and visual understanding of the evolving state of soil aridity and even the risk of desertification within these regions.

The future implementation of IoT-SOL technology for widespread soil quality monitoring across various geographical locations will enable more accurate predictions of areas exhibiting desertification tendencies.

This research will furnish critical insights into soil health, encompassing the identification of nutrient deficiencies and the presence of potentially harmful contaminants. It will facilitate the continuous monitoring of diurnal, weekly, monthly, and annual fluctuations in soil conditions. The compact design enables facile insertion into diverse substrates and soil types, permitting expeditious assessments. This, in turn, optimizes agricultural practices and empowers informed decision making to mitigate the risk of land desertification.

The resulting data can be leveraged to optimize agricultural operations, identify emerging trends, and even inform preliminary weather forecasts, ultimately maximizing yield and enhancing soil quality. The integration of IoT technologies within agriculture is widely recognized as Smart Agriculture or Precision Farming. Within this framework, the present advanced autonomous system for monitoring soil parameters emerges as a pivotal component for precision decision making.

By incorporating location-specific forecasts into desertification risk models, it becomes possible to generate more accurate predictions of future trends. These predictions can be further refined by incorporating seasonally adjusted forecasts of water availability and nutrient content within the soil profile.

7. Conclusions

This research has established foundational procedures for calibrating smart sensors and conducting valid soil quality measurements. The current system serves as a crucial initial step, representing a proof-of-concept laboratory model for the future development of an autonomous system IoT-SOL pilot station for in situ soil quality monitoring. As the first iterative system, this research focuses on the measurement and calibration of smart sensors.

Recalibration of sensors is conducted using a known soil type, emphasizing the importance of practical applicability within our research context. This necessitates rigorous testing on various known substrates. Future work will involve expanding these investigations to encompass different soil types and depths, ultimately enabling predictions of soil aridity levels.

Assessing soil quality is an evolving field intricately linked to the advancement of digital mapping and cartographic techniques, particularly in regions susceptible to desertification. Continued research in soil monitoring is crucial, necessitating the development of a long-term vision and a roadmap for technological advancements to ensure comprehensive data acquisition and analysis.

By deploying this technology across multiple geographical areas, it becomes possible to significantly enhance the efficiency of soil analysis. This enables the generation of comprehensive maps delineating regions at risk of desertification, facilitating early predictions, and enabling the implementation of targeted interventions to mitigate degradation. Ultimately, this approach contributes to expanding arable land and fostering sustainable soil management practices.

The advantages of an autonomous, multisensory soil monitoring system for assessing soil characteristics include:

Real-time data acquisition: Enables continuous and immediate monitoring of soil properties.

−

Multi-channel measurement capability: Allows for simultaneous measurement of various soil parameters, providing a comprehensive understanding of soil conditions.

−

Flexible data collection frequency: Permits adjustment of data collection intervals to optimize data acquisition while minimizing energy consumption.

−

Wireless data transmission: Facilitates efficient and convenient data transfer through the local Wi-Fi network.

−

Data export and reporting capabilities: Enables easy access, analysis, and dissemination of collected data.

This data-driven approach empowers informed decision making, optimizing the utilization of natural resources such as water and fertilizers. This, in turn, enhances crop productivity and sustainability while promoting environmentally sound agricultural practices. Crucially, this autonomous multi-sensor system for soil parameter monitoring serves as a valuable tool for identifying areas with a heightened risk of land degradation and desertification. The synergistic interplay of these advantages collectively contributes to environmental preservation.