Sustainable GIoT-Based Mangrove Monitoring System for Smart Coastal Cities with Energy Harvesting from SMFCs

Abstract

The Green Internet of Things (GIoTs) has emerged as a transformative paradigm for environmental conservation, enabling autonomous, self-sustaining sensor networks that operate without batteries and with minimal ecological footprint. This approach is especially critical for long-term mangrove monitoring in smart coastal cities, where conventional battery-powered systems are impractical due to frequent, costly, and environmentally disruptive replacements that hinder continuous data collection. This paper presents a self-sustaining GIoT sensing system for mangrove monitoring powered by sedimentary microbial fuel cells (SMFCs), enabling perpetual, battery-less, and zero-emission operation. A spatial dynamic energy management (DPM) strategy is implemented for the efficient integration of a microcontroller unit with a LoRa wireless communication transceiver and the SMFC harvested energy, ensuring a balanced self-sustained approach into a GIoT sensing network. Experimental results demonstrate an average power consumption of 190.45 μW per 14-byte data packet transmission, with each packet containing pH, electrical conductivity and ambient temperature measurements from the mangrove environment. Under the spatial DPM strategy, the network of four sensing nodes exhibited an energy consumption of 1.14 mWh. Given a harvested power density of 15.1 mW/m² from the SMFC, and utilizing a 0.1 F supercapacitor as an energy buffer, the system can support at least six consecutive data transmissions. These findings validate the feasibility of sustainable, low-power GIoT architectures for ecological monitoring.

1. Introduction

The evolution of the Green Internet of Things (GIoT) paradigm is transforming how ecosystems are monitored and protected by promoting sustainable, low-environmental impact observation strategies [1,2,3]. Built upon advances in low-power electronics, wireless communication, and eco-friendly energy harvesting, GIoT enables the development of self-sufficient sensor networks that reduce dependance on conventional energy sources and minimize ecological disturbance [4,5,6]. These innovations are particularly valuable in coastal cities, where uninterrupted data streams are essential yet difficult to sustain with traditional technologies. Mangroves, which function as natural buffers against erosion, reservoirs of biodiversity, and significant carbon sinks, play a vital role in coastal resilience [7]. Preserving their health is essential to ensure the continuity of these critical ecosystem services [8]. In this regard, real-time monitoring stands as a highly relevant tool for their conservation. However, conventional monitoring systems, often dependent on battery-powered circuits, cannot guarantee seamless, real-time monitoring, since frequent battery replacements are costly, logistically complex, and ecologically disruptive [9,10,11,12,13]. By integrating energy-harvesting technologies with intelligent sensor networks, GIoT offers a sustainable pathway for acquiring uninterrupted, high-resolution environmental data while maintaining a minimal ecological footprint [14]. The motivation for this work stems from the fundamental inapplicability of traditional batteries in these sensitive environments. Beyond the logistical burden, the risk of electrolyte leakage from damaged batteries poses a direct contamination threat to the sediment, undermining the core conservation objectives. This critical limitation compels the search for a sustainable alternative, such as sediment microbial fuel cells (SMFCs), which harvest energy in situ from the organic-rich soil, eliminate the risk of chemical pollution, and enable long-term, autonomous operation. While conventional renewable solar energy can supply plenty of electricity for IoT devices, power generation is interrupted at night [15,16]. This limitation underscores the need for alternative sources capable of continuous energy harvesting. Thus, a critical need exists for perpetual, in-situ power sources. In this context, SMFCs represent a promising energy-harvesting solution, particularly for smart coastal monitoring systems [17]. The capability for long-term, autonomous operation not only facilitates early detection of ecosystem degradation, such as erosion, biodiversity loss, or pollution events, but also supports evidence-based adaptive management strategies, thereby significantly enhancing the resilience of mangrove coasts and other vulnerable habitats. Additionally, although remote sensing technologies contribute valuable large-scale monitoring capabilities, as demonstrated in [18,19,20], they often lack the temporal or spatial resolution required for real-time decision-making. Thus, perpetual in-situ monitoring using low-power end-devices remains essential. In this context, energy harvesting from SMFCs emerges as a promising approach. SMFCs generate electricity through bacterial oxidation of organic or inorganic substrates and can be applied in organic-rich aquatic sediments. Electricity generation through electron transfer to the anode involves numerous microorganisms that form a surface biofilm, which is a result of microbial gas exchange that releases both electrons and protons [21]. SMFC efficiency relies on how well microorganisms form biofilms and their electron transfer capability. They harness the electrochemical activity of microorganisms present in the sediment of mangroves. In particular, the mangroves of Yucatan peninsula, Mexico, possess ideal conditions for evaluating the potential for energy generation through SMFCs. The high bacterial diversity, associated with nutrient inputs from aquaculture and urban discharges, could foster bacterial consortia capable of metabolizing organic and inorganic matter. This mechanism could potentially be harnessed for bioenergy generation. Furheremore, the hydrodynamics of the mangroves promote constant exchange with seawater, facilitating the renewal of microbial pools and maintaining an electrochemically active environment [22].

Studies in [23,24,25] address key limitations and challenges related to bioelectrochemical power generation, including material selection, membraneless architecture, and environmental influences. The impact of fluvial seasons on energy production is analyzed in [23]. An investigation of optimal anode and cathode configurations under constant total electrode area and identical sediment conditions is evaluated in [24]. A new power management system for multianode SMFCs is proposed in [25] to automatically disconnect impaired anodes from the rest of the system, enhancing bioturbation resilience and efficiency. However, the challengue remains to analyze the energy generation capacity in terms of the number of possible wireless transmissions enabled for GIoT systems using the harvested energy, as well as to develop a power management strategy for continuous sustainable operation.

This investigation addresses a mangrove SMFC as a natural bioelectrochemical energy source to enable a novel, self-sustaining GIoT-based network system for in-situ coastal monitoring. The proposed GIoT framework comprised of four sensing nodes and a gateway, simultaneously measures critical environmental parameters and fulfills its energy requirements autonomously through sustainable harvesting. To manage the SMFC’s variable energy output and the sensing node’s power consumption, a dynamic energy-aware strategy is implemented on each sensing node microcontroller using an edge-computing approach. Consequently, this SMFC-powered platform represents a transformative advancement toward continuous, autonomous, and ecologically integrated monitoring systems that support the preservation and restoration of sensitive coastal ecosystems. Figure 1 illustrates the conceptual GIoT-based mangrove monitoring system using energy harvesting from an SMFC.

Laboratory tests have demonstrated self-sustainable operation through a dynamic power management strategy implemented in the GIoT system for continuous coastal monitoring. This strategy automatically adjusts the wireless node duty cycle in response to variations in energy harvesting and data variability. Specifically, the duty cycle period is extended when habitat data from the mangrove remains unchanged, thereby reducing energy consumption. This technique is well-suited for self-sustaining monitoring in environments not characterized by sudden or random variations, such as mangrove coastal systems. Furthermore, an edge computing approach enables algorithm implementation directly on the microcontroller unit, eliminating the need for continuous internet connectivity to cloud-based processing. As a result, power consumption is reduced to as low as 190.45 μW per 14-byte data packet. Moreover, the spatial DPM strategy reduces energy consumption by adapting the network’s sampling rate to expected activity levels, achieving a compromise between data accuracy and power use. This approach resulted in an energy network consumption of 1.14 mWh for the four-node system. Considering a power density generation of 15.1 mW/m² from the SMFC (based on Aviccenia germinans and Rhizophora mangle), the GIoT system can perform at least 6 continuous data transmissions using an 0.1 F supercapacitor as an energy storage device.

The remainder of this article is organized as follows. Section 2 details the materials and methods used for implementing the SMFC. Section 3 describes the design of the GIoT system and the implementation of the dynamic power management strategy. Experimental results are presented in Section 4. Finally, Section 5 and Section 6 provide a detailed discussion and concluding remarks, respectively.

2. Design and Assembly of Sedimentary Microbial Fuel Cell

In this section, the design of the SMFC is described. As illustrated in Figure 2, the SMFC generates bioelectricity via an anode electrode embedded in the anaerobic sediment and a cathode electrode suspended in water.

2.1. Materials

Carbon electrodes are used for the SMFC due to their low cost, high porosity, and excellent capacity for hosting electroactive biofilms. These brush-type carbon electrodes have a high surface area that enhances cell performance by facilitating greater microbial colonization and interfacial reactions. As a result, the electrodes can generate electricity through bacterial oxidation of organic and inorganic substrates, with reported power densities reaching up to several mW/cm² [26].

In the SMFC system, electroactive bacteria at the anode oxidize organic and inorganic matter, releasing electrons that are transferred to the anode surface. These electrons flow through an external circuit to the cathode, while positive charges (protons, H⁺) simultaneously diffuse across the sediment-water interface. At the cathode, electrons, protons, and oxygen combine to form water, completing the electrical circuit. This integrated process not only generates useful energy but also contributes to the bioremediation of contaminated sediments [27].

The substrates used in MFCs also play a fundamental role by serving as a source of carbon and energy for the electroactive microorganisms in the anode, influencing their performance and affecting parameters such as power density and coulombic efficiency [26]. Among the most promising environments, mangrove sediments stand out due to their high bacterial diversity, fueled by organic matter from the sediment, algal mats, and tree debris. Figure 3 shows the locations of the collected sediment samples and the monitoring network deployed in Sisal, Yucatán, Mexico. The gateway is situated in the facilities of the National Autonomous University of Mexico, which has internet connectivity. The sensing nodes in the mangrove area are located within 500 m of the gateway.

According to Wang et al. in [28], sediments containing between 0.4% and 2.2% organic carbon provide the necessary substrate for exoelectrogens to generate electrons through respiration, which are then transferred to the anode buried in the sediments. However, the organic composition and microbiota of such sediments can vary significantly, which directly impacts electricity production. This variation is due to several factors, such as the source of organic matter (e.g., terrestrial plant debris, algal blooms, or anthropogenic waste), salinity, pH, temperature, and oxygen availability, as these conditions determine which microbial communities can thrive.

Environmental factors, such as tidal regimes and freshwater inputs, can influence physicochemical parameters [29], like salinity [30,31] and temperature [31,32]. These changes, impact conductivity, the metabolic activity of the microbial community, and ion transfer, thereby influencing the system’s energy generation and stability [29]. A previous study in Sisal, Mexico, reported a microbiota similar to that of conserved mangove sites, despite significant urban development in the area. The factors that affect the local microbiota are associated to environmental variations caused by seasonality throughout the year. Salinity, temperature, and ocean currents are the main factors that influence the structure of the microbiota [33]. This can also influence the potential for energy generation.

2.2. SMFC Fabrication

The design and construction of the SMFC prototype prioritized the integrity of electrical contacts and the correct arrangement of the electrodes. All components were constructed using corrosion-resistant materials, and connections were sealed to prevent the infiltration of water and unwanted sediments. The cell housing consisted of a cylindrical plastic container with a total height of 23 cm and an internal diameter of 12.7 cm. The container was filled with sediment to a height of 9 cm, corresponding to a volume of 1288.25 cm³ and a mass of 1571.66 g, calculated based on a wet marine sediment density of 1.22 g/cm³ [34]. Subsequently, water from the collection site was added to fill the container from the 9 cm to the 17 cm level, resulting in an additional water volume of 1.145 L.

The anode and cathode were fabricated from carbon fiber brushes and a titanium bar, respectively, each with a surface area of 214.30 cm². The electrodes were positioned 9 cm apart and connected using an insulated copper wire with a cross-sectional area of 1.5 mm².

3. GIoT-Based Monitoring System

This section evaluates the power capacity of SMFCs for designing sustainable GIoT systems that monitor mangroves in smart coastal cities. The analysis covers the sensing operations, data transmission frequency, and energy harvesting capacity of the SMFC, demonstrating the feasibility of a fully self-powered GIoT sensing node, as depicted in Figure 4.

3.1. SMFC Characterization

A SMFC was characterized to evaluate its performance under controlled environmental conditions. The assessment was conducted using aGamry Potentiostat model 1010E (Gamry Instruments, Warminster, PA, USA), which allowed for precise measurement of the cell’s electrical properties while varying key laboratory parameters.

Figure 5 presents the resulting current-voltage ($I-V$) polarization curves, which elucidate the SMFC’s performance characteristics and power generation capacity. These curves capture the cell’s response to induced environmental stress. These fluctuations could be attributed to shifts in microbial community composition, dynamic biofilm behavior, and an imbalance between anode and cathode reaction rates.

Despite these factors, the SMFC exhibited a power density variations ranging from 12.5 mW/m² to 15.1 mW/m², with a mean value of 13.8 mW/m² and a standard deviation of 1.3 mW/m². To allow for biofilm maturation on the electrodes, a 7-day period was observed prior to electrochemical analysis, as established by [35]. This variability underscores the challenge of directly powering electronics with an SMFC. To address the issue of low-level and fluctuating power output, the implementation of an energy harvester-based management system is recommended. Such a system can efficiently collect, store, and manage the variable energy from the SMFC, providing a stable power supply. Consequently, GIoT devices can maintain autonomous operation by harvesting energy directly from the SMFC, eliminating the need for an external renewable energy source or battery replacements.

The internal resistance of the SMFC was determined using electrochemical impedance spectroscopy. Measurements were performed under open-circuit voltage conditions, applying a 10 mV amplitude sinusoidal perturbation across a frequency range from 0.05 Hz to 10⁴ Hz. The internal resistance, defined as the sum of all ohmic resistances in the system including electrolyte and contact resistances, was derived from the analysis of the Bode plot generated from the impedance spectra.

Figure 6 indicates that at very low frequencies, the total series resistance is 9.37 Ω (a phase angle of 0° indicates purely resistive behavior). For frequencies of 10 Hz and above, the impedance changes, which likely arises from capacitive effects (indicated by the negative phase angle) within the SMFC’s structure.

3.2. Wireless Sensing Node Design

The RAK4630 module is a compact and versatile platform for IoT applications. It integrates a Nordic Semiconductor nRF52840 microcontroller, which features a 32-bit ARM Cortex-M4F processor (Cambridge, UK) running at 64 MHz, 1 MB of Flash memory, and 256 KB of RAM. The RAK4630 is particularly suitable for energy-constrained applications due to its multiple low-power modes, including a deep sleep mode consuming only 4.23 µA and a timer-based standby mode. The module also includes a Semtech SX1262 transceiver (Camarillo, CA, USA) compatible with LoRaWAN 1.0.3. (Beaverton, OR, USA). This combination enables the implementation of low-power, LoRa communication protocol alongside short-range connectivity via BLE, with a configurable transmit power range from −20 dBm to +4 dBm. This flexibility makes it suitable for mixed deployment scenarios. In particular, the BLE and LoRa bridge architecture offers a superior balance of range, power, and functionality compared to alternatives. Unlike power-hungry WiFi, it enables ultra-low-power operation. Compared to cellular-dependent narrow band IoT (NB-IoT), it provides longer battery life and avoids subscription fees. While Zigbee offers low power, its range is limited to personal area networks, whereas LoRa covers kilometers. Unlike the unidirectional SigFox network [36], the BLE-LoRa solution is fully bidirectional, allowing for both data uplink and device configuration. This hybrid approach leverages BLE for short-range, high-bandwidth tasks like device setup, and LoRa for efficient, long-range data backhaul, creating a versatile and cost-effective solution for distributed IoT sensing [37].

The sensing node module measures temperature, electrical conductivity, and pH, key parameters for assessing the health of the mangrove ecosystem. For sediment temperature, the module uses the SHT10 sensor. This sensor provides a 12-bit digital output with a resolution of 0.01 °C. It operates within a supply voltage range of 2.4–5.5 V, with a typical current consumption of 0.01 mA and a maximum of 1.5 μA in sleep mode. To calibrate the temperature measurements, a Ti10 Fluke infrared camera was used to acquire the temperature at soil level. The following equation is used:

$${\mathrm{T}}_{°C}=d_1+d_2·{\mathrm{DO}}_{\mathrm{T}}+d_{\mathrm{Ti}10}$$

(1)

where $d_1$ = 39.7, $d_2$ = 0.04 and $d_{\mathrm{Ti}10}$ = 0.001, which represents the calibration offset of the infrared camera. ${\mathrm{DO}}_{\mathrm{T}}$ is the digital output of the sensor.

The system also employs the low-cost SEN0169 pH sensor, which has a typical power consumption of 0.30 mA. The sensor is calibrated using standard pH buffer solutions (e.g., pH 4.0, 6.16, 9.18) in conjunction with an Ag/AgCl reference electrolyte. This calibration allows the SEN0169 to measure pH, which correlates with the concentration of soluble nutrients and chemicals in the sediment, thereby indicating their availability to mangrove plants. The following nonlinear approximation is used to adjust the pH data:

$${\mathrm{pH}}_{\left(v\right)}=-0.223v^2-4.2965v+26.832$$

(2)

where ${\mathrm{pH}}_{\left(v\right)}$ is the hydrogen potential value of the medium and v is the output voltage of the sensor. The pH measurements also require temperature compensation by applying the following correction equation:

$${\mathrm{pH}}_{comp}\left(v\right)=-0.1({\mathrm{T}}_{°C}-25°C)+{\mathrm{pH}}_{\left(v\right)}$$

(3)

where ${\mathrm{T}}_{°C}$ is the temperature at the time of measurement and ${\mathrm{pH}}_{\left(v\right)}$ is the hydrogen potential corrected for pH variation due to temperature.

The system also incorporates the low-cost SEN0114 conductivity sensor, which has a typical power consumption of 0.31 mA. Sediment conductivity ($\sigma$) plays a key role by measuring conductivity content in the sediment, a critical factor for assessing mangrove health and nutrient uptake. However, uncertainties in the measurements due to the variability of minerals and dry plant tissue make it necessary to calibrate this sensor. The following polynomial approximation is employed:

$$\sigma =11.458x+4275,\mathsf{\mu }\mathrm{S}/\mathrm{cm}$$

(4)

where x is the digital number acquired by the analog-to-digital converter available in the microcontroller.

3.3. BLE-LoRa Wireless Sensor Network Framework

Figure 7 illustrates the block diagram of the proposed wireless sensor network for monitoring mangrove parameters (temperature, pH, and electrical conductivity). Each sensing node is assigned a unique identifier and transmits data to a BLE-LoRa bridge module. The RAK4630 module supports the BLE protocol with a configurable TX power of up to +4 dBm, while its typical current consumption is 2.49 mA. After transmission, the microcontroller returns to the sleep state, to allow for the SMFC recovery effect [38]. After a BLE transmission, the bridge then relays the aggregated data to the central LoRa gateway, leveraging LoRa’s long-range capabilities. The system uses the LoRaWAN Class A protocol, which operates in the unlicensed 915 MHz band and can transmit up to 3 km line-of-sight, or up to 20 km with directional antennas. LoRaWAN technology offers an effective balance between power consumption and range coverage compared to other popular communication stacks. Configured with a spreading factor of SF7 (resulting in a 10.9 kbps bit rate, 4/5 coding rate, and 125 kHz bandwidth), the system maintains a low data loss ratio (<5%), ensuring reliable data collection for mangrove health analytics. From the gateway, sensor data is sent to a cloud network server as structured messages, using both Message Queuing Telemetry Transport (MQTT) and Representational State Transfer (REST) protocols. The processed data is visualized in real-time through an interactive web application, enabling users to monitor mangrove health parameters remotely.

3.4. Power Management Strategy

Figure 8 presents a finite state machine governing the operational states of the sensor node, designed to minimize energy consumption in the WSN through efficient scheduling and resource allocation.

According to Figure 8, the default state of each MCU sensor node is Sleep, characterized by a low current consumption of 0.5 μA. When activated with a rapid transition time of 15.6 ns, the node enters an active-sensing state to perform data acquisition and then proceeds to a processing stage for a duration of 240 ms, with a current consumption of 0.32 mA. Subsequently, the node transitions to the TX-BLE state for wireless transmission; this is the most energy-intensive phase with a peak current of up to 2.49 mA, though its duty cycle is kept minimal. Immediately after transmission, the MCU returns to the low-power sleep state, thereby taking full advantage of the SMFC recovery effect. Considering the gradual change of the Mangrove’s microbial activity and organic content during the day, a system-level DPM is proposed with spatial-temporal strategy for data acquisition. That is, we activate each node at a certain location during a specific time. Therefore, the sensor node’s energy consumption is also improved.

Assuming that the field of interest is a 2D rectangular region represented by a grid of $L\times W$ points, a deterministic deployment of the sensor nodes was implemented with a distance separating each pair of adjacent points equals one measurement unit (i.e., 1 m). The proof-of-concept network considers a number of 4 nodes deployed into the target region. Each device is placed within a square pattern at $(Li,Wi)$ for $i=1,\dots ,L;j=1,\dots ,W$, as illustrated in Figure 9, with labels $P_1$ to $P_4$.

Figure 10 illustrates the proposed spatial-temporal strategy, which employs different sensing patterns adapted to Mangrove activity levels throughout a 24-h period. The activation of sensors at specific times is indicated by color-coding. The first period, DPM-1 (red pattern), involves the sequential activation of a single sensor node every 20 min over a 12-h period. From a data analysis perspective, parameters like temperature, pH, and electrical conductivity are minimally affected at night, while the network’s power consumption is reduced by 75%. The second period, DPM-2 (black pattern), corresponds to times of low Mangrove activity during the early morning and late afternoon until sunset (a total of 4 h). This strategy activates two sensor nodes in parallel every 20 min, achieving an effective balance between measurement accuracy and energy consumption. At this stage, network power usage is reduced by 50%. The third and fourth periods represent an increase in data sampling. DPM-3 (blue pattern), applied during the morning and early afternoon, activates three adjacent sensors simultaneously (period of 4 h). Finally, DPM-4 (green pattern) engages all network sensor nodes for concurrent measurements every 20 min.

3.5. Energy Harvesting Circuit

The energy harvesting subsystem is composed of the LTC3108, a SMFC cell, and a 0.1 F supercapacitor. The LTC3108 is a highly integrated energy harvesting circuit designed to harvest and manage energy from low-voltage sources, such as a SMFC. The circuit uses a small, off-the-shelf step-up transformer and a coupling capacitor to form a resonant circuit that can boost input voltages as low as 20 mV. The circuit efficiently manages the harvested energy, even under varying input voltages. This energy is used to charge a 0.1 F supercapacitor for continuous operation, while its internal regulator provides a regulated output voltage to power a low-power sensor node. The integration of a dynamic power management allows the system to power up high-drain peripherals (like a BLE radio transmitter) only when sufficient energy is available, performs data transmission, and then shut down to resume harvesting. The LTC3108 has a typical current consumption less than 500 nA and features an integrated boost charger capable of extracting power down to the μW-level from DC energy sources. The SMFC has been characterized with an open circuit voltage of 219 mV, a short-circuit current density of 163.25 mA/m², and a maximum power density of 15.11 mW/m². The power management output has been configured to deliver a regulated voltage of VCC = 3.3 V to all the other modules of the sensor node.

4. Experimental Results

Experimental analysis was conducted using real-time measurements to analyze the energy capacity of the SMFC. Figure 11 shows a single-sensor node powered with the proposed SMFC under laboratory conditions.

4.1. SMFC Power Generation Results

The power generation capability of the system was assessed via a cold-start analysis conducted at 25 °C and under 2000 lux illumination. Figure 12 depicts the charging profile of the supercapacitor from a fully discharged state.

The LTC3108’s output voltage, shown in the figure, illustrates the initial cold-start charging phase with no initial energy stored. The SMFC successfully charged the 0.1 F supercapacitor within 90 min. According to Figure 12, the harvesting circuit attains 3.3 V in 32 min. The boost converter on the LTC3108 is configured with a VSTOR output of 5.2 V. Once the VSTOR voltage surpasses the 0.2 V threshold, the charger efficiently draws power from the SMFC. This setup allows the energy stored in the supercapacitor to power IoT sensing nodes in the absence of the primary SMFC input source. Consequently, if a BLE transmission partially discharges the supercapacitor on the VBAT terminal, a 20-min sleep period is adequate for recovery and recharge.

4.2. Sensing Node’s Energy Consumption Results

Energy consumption is evaluated for each sensor node and across the WSN. The testbed, shown in Figure 11, measures a node’s energy consumption over an operational cycle that includes active sensing and data transmission to the gateway. Communication between devices is implemented using BLE. During its operational cycle, the node first enters a sensing state, activating its sensor modules for 240 ms with a current consumption of 0.32 mA. This is followed by a transmission phase that requires a peak power of 8.24 mW at 2.49 mA. During sleep mode, the system consumes 0.59 μA. With a sleep period of 20 min, this cycle yields a total average power consumption of 190.45 μW.

Table 1. Power consumption analysis of the GIoT-based mangrove monitoring network.

Stages	Single-Sensing Node Energy Consumption per Hour	Mangrove GIoT Network Energy Consumption per Hour Spatial-DPM Strategy
Sleep + peripherals off	5.85 μWh	11.7 μWh
Active (all sensors)	3.16 mWh	6.32 mWh
BLE Tx	24.72 mWh	49.44 mWh
Average consumption	0.571 mWh	1.142 mWh

summarizes the measured energy consumption values for all operational states of the node. The Power Profiler Kit II tool of Nordic and the EnergyTrace++ tool (Dallas, TX, USA) of Code Composer Studio platform, were used for measurement and analysis of the BLE sensing node’s energy consumption.

The methodology for calculating the average network energy consumption was based on a time-weighted average of the operational modes defined by the spatial DPM strategy. Acording to the baseline energy consumption of 0.571 mWh per active node per hour, the total daily energy consumption was computed by summing the contributions from each DPM period, where the energy for a given period was the product of its network power (the number of active nodes multiplied by the baseline consumption per hour) and its duration. That is, the energy consumption per DPM period is:

• DPM-1: Energy = 1 node × baseline consumption per hour × 12 h = 6.85 mWh
• DPM-2: Energy = 2 node × baseline consumption per hour × 4 h = 4.56 mWh
• DPM-3: Energy = 3 node × baseline consumption per hour × 4 h = 6.85 mWh
• DPM-4: Energy = 4 node × baseline consumption per hour × 4 h = 9.13 mWh

This total daily energy of 27.39 mWh was then divided by 24 h achieving the final average energy consumption of 1.14 mWh for the entire four-node network, effectively modeling the energy savings achieved by adaptively scaling the number of active sensors throughout the day.

The result shows that the DPM strategy reduces the network energy consumption compared to a scenario where all nodes are always active. Thus, the energy consumption of the network per hour is 1.14 mWh. One can also assume that the energy consumption per node per hour across the network is 1.14 mWh/4 nodes = 0.285 mWh, which is 50% lower than the baseline 0.571 mWh due to the duty cycle optimization in the DPM strategy.

4.3. Sustainability Analysis and Interpretation

Based on the energy consumption analysis, the data in Figure 13 demonstrates that a fully charged 0.1 F supercapacitor can support up to six consecutive cycles of data sensing and BLE transmission without any energy replenishment from the SMFC-based harvester circuit.

This condition characterizes the worst-case scenario for the system, in which the SMFC is subjected to continuous BLE data transmissions, the most power-demanding operational mode. From Figure 12, the quantitative energy analysis demonstrates that the system successfully harvested $E_{SMFC}={\displaystyle \frac{1}{2}}\left(0.1F\right){\left(5.2\right)}^2=1.352$ J over a 90-min charging cycle. This translates to an energy generation rate of 0.901 J per hour. When compared to the system’s energy consumption, calculated as $E_{discharged}$ = (190.45 μW)(3600 s) = 0.685 J per hour under baseline operation, the SMFC harvester provides 1.31× the required energy, confirming a positive energy balance. Furthermore, by implementing our proposed DPMS during a 12-h nocturnal period, the system’s energy discharge was reduced to 0.342 J per hour. This enhancement resulted in a significantly improved support ratio of 2.64×, thereby guaranteeing system autonomy and demonstrating robust, self-sustained operation under real-world conditions. However, although the power generation capacity seems to be stable, factors such as microbial activity in the sediment, temperature fluctuations, and variations in electrolyte composition can lead to a gradual decrease in power output over several days.

To provide a benchmark, the system’s lifetime was also estimated using a traditional battery power source. A 3.7 V, 100 mAh lithium polymer (LiPo) battery, which stores approximately 1332 J of energy (0.1 Ah × 3.7 V × 3600 s/h), would power the system for approximately 2.7 months without energy harvesting. This battery lifespan was validated using the EnergyTrace++ tool within the Code Composer Studio platform, which profiles the power consumption of the microcontroller and its peripherals. In adittion, while LiPo batteries offer high specific energy density (∼140 to 200 Wh/kg) and photovoltaic (PV) panels provide high power density (∼17 W/m², for a 3 W panel), both technologies can contain toxic compounds. For instance, CdTe thin-film solar panels can contain around 5.86 g of cadmium per module, with the thin-film layer being just 1 μm thick, as reported by [15]. Therefore, although a hybrid SMFC-PV system seems an attractive energy-harvesting solution, the potential for environmental contamination from PV modules is a significant concern. This risk stands in direct opposition to the environmentally conscious principles of our GIoT approach, particularly within a sensitive ecosystem like a mangrove. In addition, the SMFC-powered system fundamentally shifts the operational paradigm from periodic, scheduled maintenance to near-zero-touch deployment.While both systems can achieve a high data yield (∼94 packets/day) and delivery ratio (98%) under good conditions, the battery-powered system suffers from high and recurring total cost of ownership (TCO) due to labor, travel, and component costs. Crucially, the environmental cost of regularly disposing of toxic batteries in a sensitive mangrove ecosystem is a significant liability. The GIoT-based monitoring system eliminates these recurring costs and environmental risks. Its TCO is dominated by the initial deployment, after which it provides stable, self-sustained data collection, making it vastly more scalable and sustainable for environmental monitoring.

To evaluate the long-term power generation capacity of the SMFC’s, an extensive field deployment was conducted over eight consecutive days to validate the operational sustainability of the mangrove sensing system. The electrochemical analysis was performed after allowing a 7-day period for biofilm maturation on the electrodes, consistent with documented formation times for comparable systems [35].

Table 2. System performance and sustainability analysis from the SMFC-based coastal system.

Parameters	Time (Days)
	1	2	3	4	5	6	7	8
SMFC electrode voltage (mV)	204	198	196	201	195	197	196	194
Average supercap voltage (V)	5.1	4.9	4.85	4.93	4.89	4.94	4.91	4.92
Supercap energy (J)	1.35	1.21	1.18	1.22	1.21	1.22	1.21	1.21
Number of Tx lossed packets per day	2	1	0	3	2	1	2	4
Packed delivery ratio (PDR) per day (%)	97.91	98.95	100	96.87	97.91	98.95	97.91	95.83
Average pH per day	8.35	8.17	8.4	8.28	8.21	8.33	8.41	8.26
Average conductivity per day (μS/cm)	134	132	135	131	129	132	130	128
Average temperature per day (°C)	26.9	28.7	26.7	27.4	28.1	27.5	26.8	27.6

presents the system’s sustainability behavior, starting from a fully charged capacitor after the cold-start process illustrated in Figure 12.

The analysis confirms a period of stable and consistent power output, which is critically dependent on the maturation of a stable electrogenic biofilm on the carbon-based electrodes of the SMFC. This biofilm is essential for efficient electron transfer and directly influences the reliability of the energy harvesting process. The analysis demonstrates a stable SMFC voltage generation range of 192 mV to 204 mV, indicating consistent bioelectricity activity. The system’s energy maintained stable voltage and energy levels throughout the entire testing period, which involved the transmission of 768 data packets over eight days (at a rate of 96 packets per day). The average energy available from the supercapacitor during this period was 1.22 J, confirming the system’s capability to support sustained operation. Regarding communication reliability, the system achieved a Packet Delivery Ratio (PDR) of 98%, with a total of 15 packets lost from the 768 transmitted. This results in a Packet Error Rate (PER) of 2%, which falls within expected margins for long-range wireless communication in challenging environments. The observed errors are primarily attributable to a combination of system configuration and environmental factors. These could be related to the gateway location and suboptimal antenna height, which impacted the link budget, as well as the selected spreading factor, which presents a trade-off between range, data rate, and airtime. Furthermore, the unique coastal conditions of the Sisal mangrove ecosystem, characterized by salt spray and high humidity, contributed to minor but periodic signal attenuation and fading, thereby influencing the overall link quality.

5. Discussion

For GIoT-driven analysis in wireless mangrove monitoring, our system enables the sustainable, self-powered sensing of sediment microbial activity, along with key parameters including pH, electrical conductivity, and temperature.

Table 3. Comparative analysis of GIoT-based mangrove monitoring networks with SMFC energy harvesting.

System Features	[17]	[18]	[19]	[24]	[25]	[26]	Our Study
Energy source	SMFC	—	—	SMFC	SMFC	SMFC	SMFC
Monitoring system	Sensor network	SPOT-5	IKONOS, LIDAR	—	—	—	GIoT
Power generation	3.4 mW/m2	—	—	394 μW/m2	32 mW/m2	49 mW/m2	15.1 mW/m2
Sensor node consumption	2.5 W	—	—	—	—	—	1.45 mW
Communication protocol	—	—	—	—	—	—	BLE-LoRa
Mangrove analysis	—	Indirect (NDVI, EVI)	Mangrove type classification	—	—	—	Mangrove health (Temp, pH, EC)

presents a comparative analysis of the power capacity of SMFCs for supporting such a GIoT-based system against other similar works.

Previous studies have developed remote sensing for mangrove monitoring [18,19] and energy harvesting from Sediment Microbial Fuel Cells (SMFCs) [17,24,25,26]. For instance, research in [18,19] has proposed mangrove mapping and monitoring using remote sensing techniques to enhance climate change resilience and knowledge visualization. Although these studies classify mangrove types using NDVI and EVI from satellites like SPOT-5 and IKONOS, their spatial resolution and data interpretation methods are insufficient for real-time monitoring of mangrove health.

On the other hand, recent works [17,24,25,26] have demonstrated the ability to harvest energy from SMFCs. While this energy harvesting approach is promising, reporting power outputs of 32 mW/m² and 49 mW/m², higher than our system, these studies employ an array of SMFCs and did not incorporate a GIoT system. Such integration is necessary to sustainably monitor mangrove dynamics.

Moreover, seasonal environmental variations, with salinity, temperature, and ocean currents being the primary factors, drive changes in the microbiota structure [33]. These microbial shifts directly influence the potential for energy generation in the mangrove. A Dynamic Power Management (DPM) strategy is proposed to establish an adaptive balance between the mangrove monitoring duties and the power requirements of the sensing system. This DPM methodology considers the dynamic changes in the sediment’s microbial activity and organic content throughout the day and can be adapted to the soil features of different locations. In this regard, artificial intelligence agents could be integrated as a novel alternative for the GIoT monitoring system. However, the practical implementation of AI agents introduces its own set of challenges, including the computational overhead required for on-device learning and the need for extensive, site-specific data for training, which may not be feasible for all deployment scenarios.

6. Conclusions

This study presents a sustainable GIoT-based monitoring system for smart coastal cities, entirely powered by SMFCs. The core of this system is a custom-designed SMFC prototype with carbon brush electrodes, rigorously characterized to validate its reliable power generation for the sensing nodes. The implemented sensing node, architected around the low-power nRF52840 microcontroller, achieves a critical ultra-low power consumption of 190.45 µW. A spatial DPM strategy further enhances real-world viability by intelligently aligning data sampling with microbial activity cycles, optimizing the trade-off between data fidelity and power autonomy. The entire four-node network consumes 1.14 mWh, underscoring its operational efficiency. This low-energy architecture is successfully matched with an LTC3108-based energy harvester, which interfaces with the SMFC to extract a maximum power density of 15.1 mW/m². The resulting positive energy balance, a key finding of this work, enables up to six consecutive BLE transmissions, ensuring reliable data communication in a field setting. Ultimately, the integration of SMFC energy harvesting with an ultra-low-power platform presents a viable, self-sustaining solution for protecting vulnerable mangrove ecosystems. The profound ecological significance of this work lies in its non-invasive operation, which eliminates the risk of battery-derived chemical contamination and establishes a truly symbiotic relationship with the environment it monitors. By enabling perpetual, in-situ monitoring of critical parameters, this system provides a practical technological foundation for the early detection of erosion, pollution, and biodiversity loss, thereby directly contributing to the preservation and resilience of these vital coastal buffers.