Designing CO₂ Monitoring System for Agricultural Land Utilizing Non-Dispersive Infrared (NDIR) Sensors for Citizen Scientists

Abstract

The increasing atmospheric CO₂ concentration due to anthropogenic activities has led to the development of low-cost, portable, and user-friendly sensing technologies. Non-Dispersive Infrared (NDIR) sensors offer reliable CO₂ detection with high sensitivity, which makes them ideal for citizen scientists. In this context, we designed two low-cost CO₂ monitoring systems: an automatic opening chamber with a lid and a portable device using NDIR sensors. These monitoring systems were calibrated (R² = 0.99) with known CO₂ concentrations. Besides its reliability and accuracy, the Automated CO₂ Monitoring System costs approximately USD 220.77 and portable CO₂ device costs USD 151.43, which makes them suitable for citizen scientists. Due to CO₂ gas monitoring system’s simplicity, structure, and operation, non-expert users can use and actively participate in environmental monitoring data collection. This promotes public engagement in climate and air quality monitoring and enables citizen scientists to have reliable data for CO₂ monitoring and environmental awareness.

1. Introduction

A continued pressing concern around the world is increased greenhouse gas emissions into the atmosphere [1]. However, there are many sources from which these gases are emitted. One that may be overlooked easily is the release of greenhouse gases from farmland [2]. The process of irrigating farmland has various components that contribute to global greenhouse gas emissions [3]. The ground is the second largest carbon sink [4], right after the ocean, and is an important part of the carbon cycle. Usually, soil acts as a sink through carbon sequestration [5], but when lands, such as forests or grasslands, are converted into agricultural fields, they can instead become a source of carbon in the cycle [4]. Microorganisms exist in the soil and when water suddenly flows in, their respiration rate increases, releasing carbon dioxide and other greenhouse gases into the atmosphere [6]. When land is transformed into agricultural fields, this process can then be further amplified due to the decreased carbon sequestration that usually offsets this respiration from soil microbes. Efficient irrigation practices can help offset these greenhouse gas emissions by limiting excess water consumption [7]. Although CO₂ emissions from agriculture have been a known contributor, how these emissions are further broken down is where information can become sparse. Extensive previous research has been conducted on the farming equipment’s contribution to CO₂ emissions, or the pumps required to irrigate the fields, as well as emissions of greenhouse gases from the soil [8,9]. Understanding emissions from agricultural land is important to implement better agricultural practices to limit and reduce the environmental impacts. A previous study has shown that cover crops and mulching can reduce these greenhouse gas emissions from soil [10]. Another study evaluated the effect of irrigation methods on soil greenhouse gas emissions and stated that soil moisture and temperature are key factors affecting greenhouse gas emissions [11].

CO₂ monitoring systems are valuable tools in agricultural operations, which offer understanding about crop health, soil condition, and environmental impact. Recent research has focused on innovative sensing technologies for CO₂ monitoring in agriculture. For grain quality issues during storage, researchers have developed a continuous CO₂ monitoring system [12]. For soil CO₂ assessment, the Internet of Things (IoT)-based CO₂ sensor probes are used to measure CO₂ concentrations [13]. An Arduino-based monitoring platform for atmospheric, soil, and dissolved CO₂ concentrations has been developed by researchers [14]. The different sensing technologies include NDIR sensors, electric field sensors, and Tunable Diode Laser Absorption Spectroscopy (TDLAS) [15]. Currently, there is not a large range of technology on the market to support greenhouse gas emission research or data collection [16]. Static gas collection chambers are current and reliable methods for monitoring CO₂ flux within agricultural applications, with systems like the eosAC-LT/LO Automated Soil Flux Chamber and LI-COR Smart Chamber allowing for measurement, scheduling, and continuous soil gas flux monitoring. The high cost of these systems limits their use by citizen scientists (farmers, field staff, agricultural professionals, etc.). With growing interest from citizen scientists in monitoring soil gas flux, the development of affordable CO₂ gas monitoring systems is needed.

This study introduces and designs a portable CO₂ monitoring device and an automatic opening chamber CO₂ monitoring system with unique designs and affordable prices compared to commercially available systems. Thus, the purpose of this study is to provide the design and development of an affordable and user-friendly CO₂ monitoring system for citizen scientists using microcontrollers and IoT (Internet of Things) technology.

2. Materials and Methods

This study focused on the development and design of a low-cost and innovative CO₂ monitoring system using Non-Dispersive Infrared (NDIR) sensor technology for citizen scientists. Two different systems were designed to address CO₂ monitoring. One is an automatic opening chamber, and the other is a portable device. The automatic opening chamber is designed for stationary and precise CO₂ flux measurement, while the portable device is primarily designed for flexible and on-time CO₂ flux monitoring.

2.1. Design 1: Automated CO₂ Monitoring System

An automatic opening chamber was designed to monitor precise and uninterrupted CO₂ flux in diverse environmental conditions. To minimize manual interruptions, external interventions, and consistent sampling, a motorized lid was designed for the chamber that opens and closes automatically at programmed intervals. The system runs on a 30-min loop, with the chamber being in the closed position for 4 min and then opening for 26 min before repeating the loop. AutoCAD version 24.1 was used for custom fitting parts of the chamber.

2.1.1. Hardware Design

The automatic opening chamber comprises a lid, worm gear motor, and a cylindrical chamber. The lid and chamber were both designed in AutoCAD version 24.1 and printed using a Bambu Lab P1S 3D printer (Bambu Lab., Shenzhen, China) using PETG green filament. The materials used for printing are durable and resistant to varying environmental conditions. The automated lid was connected to a worm gear motor (single shaft, 12 volts with 110 RPM manufactured by Greartisan) and controlled with an Arduino microcontroller motor driver (HiLetGo, Shenzhen, China). A rechargeable 12-volt power lead acid battery (Expert Power, Paramount, CA, USA) with 2.1 Ampere output was used for power supply to the motor. The NDIR CO₂ sensor was adjusted in the chamber for accurate CO₂ flux measurement. Figure 1 shows the system being used in a blueberry research field at Michigan State University’s south campus farms (East Lansing, MI, USA). The individual components that make up this system are displayed in (Figure 2).

2.1.2. Software Design

The software design of the automatic opening chamber CO₂ system was built on an open-source platform using two separate Arduino boards. One ran the code for the CO₂ sensor and micro-SD card writer, and another monitors the motor’s functions. The CO₂ sensor’s code was initially taken from GitHub and further tailored for CO₂ measurement. The GitHub code only ran the CO₂ sensor but was unable to save the data to an SD card, which was altered to enable the data storage function. The opening and closing of the chamber’s lid were controlled by code, with options for how long the loop function performs its speed, and duration that the motor spins either clockwise or counterclockwise. By integrating the components, a 30-min loop function was created, where the first 4 min involved in recording and saving data while the lid is closed, and the next 26 min were used for recording and saving data with the lid open. Figure 3 shows a flowchart of the code for the operation of the chamber. The code comprises four input commands, where typing “A” into the serial monitor rotates the motor clockwise, while typing “C” rotates the motor counterclockwise for 500 milliseconds. For the codes to work, the lid needs to start in an “open” position, which can be manipulated with the command’s “A” and “C”. Giving the command “Go” to the serial monitor causes the code to begin its loop, with the initial step being the closing of the lid. After 4 min, the lid opens, and data is further collected for 26 more minutes before the code loops. The “Stop” command stops the code from running and requires the user to input another “Go” before the loop restarts. At the beginning of each loop, the code creates a new file on the SD card to be able to separate data for separate trials.

2.2. Design 2: IoT-Based Portable CO₂ Monitoring System

2.2.1. Hardware Design

The schematic workflow of the portable CO₂ monitoring system (Figure 4) prioritizes space, efficiency, portability, and ease of assembly due to durable housing, modular electronics, and a custom-designed sensor chamber. The system’s components are housed in a Polycase ML-57F (Polycase, Avon, OH, USA) enclosure with an ML-58K aluminum mounting plate (Polycase, Avon, OH, USA). The Polycase enclosure has an Ingress Protection of 68 (IP68), which does not permit dust and water to enter. Its polycarbonate walls are easy to modify for custom fittings. For this experiment, a 6.35 mm cable gland was installed at the enclosure’s bottom to connect the CO₂ sensor.

Inside the enclosure, components are mounted on the aluminum plate, as shown in Figure 5. Four holes were drilled to secure an Adafruit Quad 2 × 2 FeatherWing Kit (Adafruit, Brooklyn, NY, USA) using M 2.5 standoffs and screws. The kit supports up to four modules. In this setup, it holds an Adafruit FeatherWing OLED 128 × 64 display module and a Particle Boron LTE (Particle, San Francisco, CA, USA) 4G LTE cellular microcontroller. The collected data is sent to the Ubidots IoT web server. The Sensirion CO₂ sensor is connected to the microcontroller via a QWIIC cable.

The system is powered by a Voltaic V50 USB battery pack (Voltaic, Brooklyn, NY, USA), which features two USB Type-A and two USB Type-C ports. Its 12,800 mAh capacity ensures reliable operation of the microcontroller and sensors during data collection, making it ideal for IoT applications. The Voltaic battery packs can be integrated with solar panel if extended power is needed for the collection of CO₂ data.

To optimize CO₂ data collection, a custom chamber was designed in AutoFusion CAD. Measuring 180 × 93 mm, it was 3D printed with white PLA filament using a Bambu Lab P1S printer (Bambu Lab., Shenzhen, China). The chamber aligns with the enclosure’s bottom, mounting easily with four bolts and nuts. To ensure an airtight seal, 6.33 × 1.58 mm weather-stripping was applied between the chamber and enclosure.

2.2.2. Software Design

The software for the portable CO₂ monitoring system was designed to ensure accurate data collection, efficient data transmission, and user-friendly operation (Figure 6). It combines firmware development, cloud services, and data visualization for the seamless integration of hardware components and real-time functionality. The particle microcontrollers were programmed using Particle’s IDE, either through the Web IDE or the Workbench in Visual Studio Code version 1.86. The codes for this system were adapted from example files provided by the library dependencies of the components. The Adafruit FeatherWing OLED display provides a live visual interface for the system. The display presents real-time CO₂ readings, as well as info about data transmission status or any errors that may arise. To conserve energy, the Particle Boron enters low-power sleep mode during idle periods between readings. During standby, the OLED and sensor are completely powered down to minimize power loss. Particle Boron manages the wake cycle, ensuring the system resumes operation at the programmed intervals (every 10 s for 5 min) without user intervention.

2.3. SCD 30 Sensor Calibration and Testing

Prior to data collection, the Senirons SCD 30 sensor was calibrated. First, the system was tested with a leakproof smoke testing kit. The smoke kit was attached to the intake tubing and allowed to run to see the flow of gas throughout the system to identify any leaks. This allowed testing of the Sensiron AG SCD30 in a controlled environment, without the entrance of other gases, using a small clear plastic chamber connected to Gasco CO₂ cylinders of varying mg/L values, from 500 to 2000 mg/L (Gasco Affiliates, LLC., Oldsmar, FL, USA). Each Sensiron AG SCD30 was calibrated to ±2% of the desired concentration and balanced with oxygen. Each cylinder was pressurized to 1000 psi prior to purchase and had valve openings of 15.9 mm. To connect the gas cylinders to the calibration chamber, the cylinders were fitted with a pressure regulator, which was attached to the chamber via rubber tubing. The chamber had three openings, one each for the sensor’s wires, the gas inlet, and the gas to escape the chamber. Figure 7 shows the setup for the sensor calibration. The dimensions of the chamber length, width, and height were 88, 64.1, and 56.4 mm, respectively. The CO₂ sensor is connected to an Arduino Uno R3 (Arduino, Somerville, MA, USA) and runs via the Arduino desktop application, using hand-tailored code made to collect data every 5 min before repeating. To access the collected data, an Adafruit 5v Ready Micro-SD Breakout Board+ (Adafruit, New York, NY, USA) was used and connected to the Arduino. To power the setup, the Arduino Uno R3 simply had to be connected to a computer or laptop via a USB type B cable.

Using custom code, the sensor collected CO₂, humidity, and temperature at every 5 s for 5 min (300 s). For calibration, only the CO₂ data was used, which was saved to a Micro Center 16 GB SD card (Micro Center, Hillard, OH, USA), and transferred to Excel via a .txt file for further statistical manipulation. For stabilizing the chamber to measure the cylinder’s concentration level, the chamber was left for 15 min prior to data collection. Data was collected three times for CO₂ concentrations ranging from 500 to 2000 mg/L. The programming was adjusted by adding 104.82 mg/L to each sensor value to ensure accurate reading.

The CO₂ flux (F_c; mole m⁻² s⁻²), leaving the surface, was calculated through Equation (1):

$$F_c={\displaystyle \frac{V{P}_o}{RS{T}_o}}·{\displaystyle \frac{dc}{dt}}$$

(1)

where F_c refers to the CO₂ flux, R is the gas constant number 8.314 (joule kelvin⁻¹ mole⁻¹), V is the volume of the air (liters) exposed above the surface area of the soil face within the jar, Po indicates the pressure in the container (assumed to be atmospheric in pascals), and S represents the surface area being measured for flux, which was (38.465 cm²), as the face of exposed soil within the cylinder. The $T_o$ represents the temperature (assumed to be 25 °C, as the experiment was conducted in a temperature-controlled environment) and dc/dt represents the initial rate of change in the CO₂ mole fraction. The change in CO₂ over time for post-irrigation (5 min) was calculated.

Raw CO₂ curves were normalized to create delta values using the initial faster rate followed by the slower rate over time. Following the normalization of curves, the delta CO₂ was computed from the maximum value from the normalized figure minus the minimum value. By omitting the first stage of CO₂ measurement, the more stable second stage provided a better comparison between treatments, as it persists for longer periods and gives a more accurate representation of the change in CO₂.

After calibration, the Seniron was used in a temperature-controlled condition laboratory for testing. The treatments comprised two moisture contents (16 and 32%) along with a control (no water application), replicated five times (Figure 8). About 9 cylindrical plastic jars with a volume of 706.86 cm³ were filled with 280 g of dried sandy loam soil collected from Michigan State University’s south farm. The CO₂ data of each experimental unit was recorded for ten minutes, with five minutes prior to irrigation and five minutes after irrigation, allowing time in between to drain the applied water fully. The data was collected for each treatment and replication.

2.4. SCD 30 Performance Field Evaluation

The Sensiron AG SCD30 was deployed in an agricultural field and recorded CO₂ concentration data for about 50 days. The aim was to assess the sensor’s CO₂ concentration measurement relationship with environmental factors, especially temperature (°C), humidity (%), and rainfall (mm). The sensor recorded a time series (every 30 min) of the CO₂ concentration, and the data were sent to the IoT server using an embedded 4G LTE modem. The SCD30 sensor was in the WH31-SRS solar radiation shield to avoid getting wet.

2.5. Statistical Analysis

The CO₂ sensor’s performance and accuracy were evaluated through calibration. The calibration curve was generated, and the coefficient of determination (R²) was used to assess the sensor’s measurement reliability. The CO₂ change and flux were calculated using formulas. The data obtained were presented as mean values, with error bars indicating the standard deviation to reflect measurement variability.

3. Results and Discussion

3.1. SCD30 Sensor Calibration

The CO₂ sensor was calibrated through comparing sensor readings with corresponding values of known CO₂ concentrations. The collected CO₂ (mg/L) sensor values were plotted against the known CO₂ concentrations (mg/L), as shown in Figure 9. The relationship between the sensor value and the known CO₂ concentration was assessed with a linear trendline, as shown in Figure 9. The regression equation showed a strong linear relationship with R² of 0.999, showing the high accuracy and reliability of the sensor’s CO₂ measurement. Other commercially available sensors, such as Sunrise AB and K30, exhibit an accuracy of ±3%, while the GMP343 probe has an accuracy of ±1% [17].

3.2. Design 1: Automated CO₂ Monitoring System

This Automated CO₂ Monitoring System comprises a motorized lid mechanism operated by codes and a microcontroller (e.g., Arduino), which enables lid opening and closing automatically at programmed intervals. The chamber materials are durable enough to cope with environmental stresses in open fields. This is an energy-efficient and user-friendly automatic system for citizen scientists. This system is low-cost for building and installation, compared to current CO₂ monitoring systems that are commercially available. The components, including quantity, cost per unit, and the manufacturer of the products for making an Automated CO₂ Monitoring System, are presented in

Table 1. Billed list used for Automated CO₂ Monitoring System.

Components	Quantity	Cost Per Unit (USD)	Manufacturer
Aluminum Extrusion	1 pk	9.99	VXB (Lake Forest, CA, USA)
Arduino Uno R3	1	27.6	Arduino (Somerville, MA, USA)
Adafruit SCD-30 CO2 Sensor	1	58.95	Adafruit (New York, NY, USA)
L298N 12v Motor Controller	1	6.49	DIYmall (Shenzhen, China)
Adafruit 5v MicroSD Card Breakout Board	1	7.5	Adafruit (New York, NY, USA)
Greartisan 12v 110 RMP Wormgear Motor	1	14.99	Greartisan (Shenzhen, China)
PETG Green Filament	1 roll	17.99	OVERTURE (Shenzhen, China)
12v 7A Battery	1	19.99	Mighty Max Battery (New York, NY, USA)
Small Plastic Container	1	14.99	Superio (New Jersey, USA)
SoilWatch 3.0 Soil Moisture Sensor	1	18	Pino-Tech (Stargard Zachodniopomorskie, Poland)
Jumper Wires	75 pk	4.95	Adafruit (New York, NY, USA)
16GB Micro SD Card	1	6.19	SanDisk (Milpitas, CA, USA)
USB A Male to USB B Male	1	7.15	Amazon Basics (Seattle, WA, USA)
12v Alligator Clips	3 pk	5.99	QTEATAK (Shenzhen, China)
Total cost (USD)		220.77

. The eosAC-LT/LO Automated Soil Flux Chamber, initially introduced within the Introduction, is a currently marketed system capable of remotely recording and monitoring CO₂ and other gas emissions from the soil. This system, however, depending on the quote, has ranged from USD 5000 to 10,000. These quotes do not include the electricity to run the systems and a computer, which it is assumed the prospective citizen scientist already has access to. Our Automated CO₂ Monitoring System costs approximately USD 220.77 (Table 1). Other CO₂ monitoring equipment is on the market, yet they lack the control chamber included in our Automated CO₂ Monitoring System or the eosAC-LT/LO system. Other published systems that include a mount like the one described in our paper are significantly more expensive [18]. This includes a system that costs USD 385.00 with a mounted attachment. The Vernier CO₂ Gas Sensor (Vernier, Beaverton, OR, USA) is capable of reading CO₂ mg/L levels of 0–10,000 or 0–100,000. This range is higher, but greatly exceeds the range needed for soil CO₂ monitoring. Its unit cost, however, is USD 299.00 and does not include a chamber, which makes it more expensive than our Automated CO₂ Monitoring System (USD 220.77).

3.3. Design 2: IoT-Based Portable CO₂ Monitoring System

The IoT-based portable CO₂ monitoring system equipped with an NDIR sensor is designed for mobility and is easy to use for real-time data monitoring. It is housed in a compact, weather-resistant, lightweight, and durable enclosure, powered by a power bank and supports data storage and remote data access. The components, unit price, and total price of an IoT-based portable CO₂ monitoring system costs approximately USD 151.43 (

Table 2. Billed list used for the portable CO₂ monitoring design.

Components	Quantity	Cost Per Unit (USD)	Manufacturer
Adafruit FeatherWing OLED—128 × 64	1	14.95	Adafruit Product ID: 4650 (New York, NY, USA)
ML-57F Weatherproof	1	31.59	Polycase (Avon, OH, USA)
Particle Boron	1	48.95	Particle (San Francisco, CA, USA)
Sensiron AG SCD30	1	19.38	Sensirion AG (Zurich, Switzerland)
Jumper Wires	75 pk	4.95	Adafruit (New York, NY, USA)
INIU Portable Charger, Slimmest 10,000 mAh 5V/3A Power Bank	1	19.99	INIU (Shenzhen, China)
Half Sized Premium Breadboard—400 Tie Points	1	4.95	Ada Fruit (New York, NY, USA)
MicroUSB	1	7.12	Amazon Basics (Seattle, WA, USA)
Total cost (USD)		151.43

). Comparable articles advocating low-cost sensor systems do not include an aluminum capturing device; rather, they include an enclosure system with no means of mounting to the ground or a station. An example is provided by [19], with USD 215 as a recent low-cost option.

3.4. Laboratory Experiment

The CO₂ patterns observed from the IoT server Ubidots (Medillllen, Columbia), following the wet treatment, followed an immediate pulse pattern in which there was a peak CO₂ reading within the first 20% of the time collected, followed by a dramatic stabilization, as shown in Figure 10. A sresearch described that CO₂ concentrations initially increase at a faster rate and then remain stable [20]. Two promising rates are observed in this trial, showing the initial linear increase with a faster rate followed by a slower rate and then remaining stable. The first faster rate experienced a change from the atmospheric CO₂ mg/L to a peak, with an average duration of two minutes. The second slow-rate CO₂ mg/L measurement typically followed the last three minutes of the experiment and presented a much more stable reading. Following the normalization of the curves, the delta CO₂ was computed from the maximum value from the normalized figure minus the minimum value. By omitting the first stage of CO₂ measurement, the more stable second stage provided a better comparison between treatments, as it persists for longer periods and gives a more accurate representation of the change in CO₂.

The bar graph (Figure 11) depicts the measured CO₂ flux with the Sensiron AG SCD30 in different soil moisture contents, along with the control (dry soil). This figure showed that wet soil produced more CO₂ than dry soil. It is clear that higher soil moisture content increases CO₂ flux. The maximum CO₂ flux was observed for soil with 32% moisture content, followed by soil with 16% moisture content. The dry soil showed negligible CO₂ flux. This figure suggests a positive correlation between the moisture content and the CO₂ flux over time. Research demonstrated that CO₂ flux in drying and ambient soil is less than the CO₂ flux in wet soil [21]. A rapid increase in CO₂ flux at the start of a wetting event is reported [22]. An increase in soil moisture results in an increase in CO₂ concentration with a peak and decline pattern for CO² emissions at 16% water-filled pore space [21]. Another study reported a positive correlation between soil moisture contents and the CO₂ concentration [23].

3.5. CO₂ Concentration Monitoring in an Agricultural Field

Outside of a laboratory environment, the Sensiron AG SCD30 was deployed for field testing (50 days) to continuously monitor the CO₂. The field deployment of the CO₂ monitoring system indicates its capability to capture CO₂ dynamics over time alongside environmental factors (temperature, humidity, and rainfall), as shown in Figure 12. The data revealed that CO₂ exhibits a moderate to strong negative correlation with temperature with a Pearson correlation coefficient (r = −0.594). This suggests that as the temperature increases, CO₂ decreases due to the influence of temperature on photosynthesis. A moderate positive correlation of CO₂ with humidity (r = 0.528) suggests that an increase in humidity also increases CO₂ concentrations, while CO₂ dynamics did not show any significant linear relationship with daily rainfall. A negative correlation between the temperature and CO₂ was also reported by Zhu et al. [24], while a positive correlation between humidity and CO₂ was reported by Singh and Malarvili, [25]. A strong documented correlation exists between the CO₂ concentration and temperature [25]. The data collected within the field evaluation in Figure 12 documents the positive trend confirmed at our test site with the Sensiron AG SCD30. For the CO₂ monitoring system kinetics improvement, the selection of an optimized sensor that compensates with environmental fluctuation, drift prevention, advanced algorithms, machine learning, and predictive modeling can be used for sensor performance. The narrowband optical filtering makes the NDIR sensor highly specific to CO₂, even in a mixture of different gases [26]. The figure indicates that individual factors did not fully explain the variations, indicating the need to account for other environmental and biological factors in CO₂ assessment. This system demonstrates the potential to monitor CO₂ continuously, providing avenues to assess complex interactions between CO₂ and its influencing factors. This system is beneficial for long-term environmental monitoring research studies and citizen scientist use.

4. Conclusions

The CO₂ monitoring systems designed in this study, including an Automatic Opening Chamber and a portable CO₂ monitoring device, were tested for their reliability, accuracy, and continuous data measurement. Both systems are low-cost compared to the commercial systems available for CO₂ monitoring. NDIR sensors were used and calibrated in both systems. These systems are accurate, reliable, and low-cost, which makes them suitable for citizen scientists. The Automatic Opening Chamber CO₂ monitoring unit cost is USD 220.77, and the portable CO₂ monitoring system unit cost is USD 151.43. Due to these CO₂ gas monitoring systems’ simplicity, structure, and operation, non-expert users can use them and actively participate in environmental monitoring data collection. The Automated CO₂ Monitoring System provides a valuable solution for continuous CO₂ monitoring over time, while the IoT-based portable CO₂ device offers mobility, ease of use, and real-time measurement. Both systems are cost-effective and accurate and are designed to help and empower citizen scientists to contribute to agricultural and environmental research. Laboratory testing shows sensor accuracy and reliability in capturing CO₂ dynamics under different soil moisture conditions. The field deployment of sensors shows the systems’ ability to capture CO₂ fluctuation and trends over time under different environmental factors. However, there is a need for long-term field sensor deployment research to assess the sensors’ degradation and drift for sensor accuracy and reliability. These systems offer practical, user-friendly, and cost-effective solutions for citizen scientists. They promote public engagement in climate and air quality monitoring, which enables citizen scientists to obtain reliable data for CO₂ monitoring and environmental awareness.