Sesamum indicum L. (sesame) is a significant oilseed crop known for its high-quality oil rich antioxidants and essential fatty acids, with considerable nutritional, medicinal, and economic value. In June 2023, sesame plants at the Agricultural Experimental Farm, Calcutta University, Baruipur, West Bengal, India (22°22′ N, 88°26′ E), exhibited chlorosis, wilting, and root rot, affecting approximately 56% of the plants, distributed randomly across the field. Initial symptoms included drooping leaves, chlorosis, and wilting, leading to plant death while foliage remained intact. Cross-sectional analysis of the collar region revealed necrotic tissues and dark brown discoloration in vascular and cortical tissues, with dark brown microsclerotia on stem bases and blackened roots. Pathogen isolation was performed using root and crown tissues from five symptomatic plants. Tissue sections (0.5–1 cm) were surface-sterilized with 2% NaOCl for 2 minutes, rinsed with sterile water, and plated onto potato dextrose agar (PDA). The isolates displayed rapid mycelial growth at 28°C under a 12-hour photoperiod. Initially, the mycelia were hyaline (average 4.13 μm width), later turning grey to black within 7 days. Spherical to oblong microsclerotia (75 μm × 134 μm on average) developed within 5 days. For molecular identification, genomic DNA was extracted from a representative isolate (VAS10). Amplification of ITS, TEF-1α, CAL, ACT and β-TUB regions was conducted using primers ITS1/ITS4 (White et al., 1990), EF-728F/EF-986R, CAL-228F/CAL-737R and ACT-512F/ACT-783R from Carbone and Kohn, 1999 and T1/T22 (O'Donnell and Cigelnik, 1997), respectively. The resulted amplicons were sequenced at Barcode Biosciences, Bangalore, India and deposited in the NCBI GenBank database (ITS: PQ368303, TEF-1α: PQ383497, CAL: PQ383499, β-TUB: PQ415078, and ACT: PQ383498. Sequences were aligned with several isolates of Macrophomina tecta (MK968306, MW592218, MW592136) previously reported (Poudel et al., 2021) using ClustalW. The pathogenicity of M. tecta was tested on 6-week-old sesame genotypes (var. Rama and VRI-1) planted in 8-inch pots filled with autoclaved soil and maintained in a greenhouse at 28±3°C with 75% RH. The inoculum was prepared by culturing M. tecta in potato dextrose broth for 7 days at 28°C. Mycelial mats were blended in 250 mL of sterile distilled water, filtered through four layers of cheesecloth, and adjusted to a concentration of 10⁵ microsclerotia/mL using a hemocytometer. Fifteen plants per genotype were inoculated via soil drenching with 50 mL of the suspension; while five control plants received sterile water. After 3 weeks, inoculated plants exhibited lower stem lesions and microsclerotia formation, whereas control plants remained healthy. The pathogen was reisolated from infected plants and identified as M. tecta based on morphological and molecular analysis. Koch’s postulates were confirmed through two independent repetitions with consistent results. Previously, M. tecta was reported on sorghum and mungbean in Australia in 2019 by Poudel et al. This study represents probably the first report of M. tecta infecting sesame, both in India and globally. The pathogen thrives in high humidity and elevated soil temperatures, posing a significant threat to sesame production in India. The emergence of this new species underscores the need for further research and management strategies to mitigate its impact on sesame cultivation.

Ground-dwelling insects are key indicators of ecosystem health due to their roles in decomposition, nutrient cycling, and the dynamics of food webs. Post-mining lands often suffer from environmental degradation, requiring effective reclamation strategies. This study assessed the relationship between microclimate and ground-dwelling insect composition in reclaimed limestone and clay mining sites at PT Semen Tonasa, Pangkep Regency, South Sulawesi, Indonesia. Insects were sampled using pitfall traps across sites with different reclamation years. Diversity indices Shannon-Wiener (H'), richness (R), evenness (E), dominance (D), correlation analysis, and principal component analysis (PCA) were used to evaluate patterns. A total of 23,294 individuals representing 36 species, 26 families, and nine orders were recorded. Dominant species included Dolichoderus thoracicus and Carebara diversa, indicating high ecological adaptability. The highest diversity (H' = 2.09) and richness (R = 3.32) were found in the 2017 clay site, while the 2018 limestone site had the highest evenness (E = 0.85) and dominance (D = 0.36). Correlation analysis showed that soil temperature and humidity were significantly positively associated with insect diversity (r>0.60, p<0.05). PCA revealed air temperature, humidity, soil pH, and light intensity as key factors influencing insect communities, accounting for 52.08% (PC1) and 28.63% (PC2) of the variance. These findings highlight the importance of microclimate-informed management for successful post-mining land restoration.

Urban development has caused significant vegetation loss, leading to air pollution, increased temperatures, and ecological imbalance. Vertical gardening addresses space constraints in cities, but manual maintenance is inefficient and costly. This work presents an IoT-based intelligent vertical garden system using a NodeMCU microcontroller and environmental sensors to continuously monitor plant conditions and automatically control irrigation. A 30-day field study showed 44% water savings, 90% cost reduction, and 98.7% system reliability, making it suitable for residential, commercial, and public spaces. The system is supported by an Android application that enables remote monitoring and control. Users can view real-time soil moisture, temperature, humidity, and pH values, switch between automatic and manual modes, adjust thresholds, and initiate irrigation instantly. The app also provides alerts for low moisture, system failure, or abnormal behavior, ensuring timely intervention. This integrated hardware-software approach makes automated vertical gardening practical for homes, hostels, and small institutions

Traditional irrigation practices in banana cultivation often result in inefficient water use, reduced productivity and higher environmental costs. With increasing challenges posed by climate variability and resource scarcity, smart irrigation technologies have emerged as a sustainable solution. This review article synthesises current research on the integration of Internet of Things (IoT), Machine Learning (ML) and Deep Learning (DL) in banana irrigation management. IoT sensors, including soil moisture, temperature, humidity and flow meters, enable real-time data collection and precise water delivery systems. ML algorithms such as regression models, random forests and Support Vector Machines (SVMs) support predictive irrigation scheduling and water requirement forecasting. DL techniques, particularly Convolutional Neural Networks (CNNs), are increasingly applied in image-based monitoring for detecting water stress and disease, integrated with drones and satellite imagery. IoT–ML–DL frameworks, supported by cloud computing and mobile applications, that create automated, data-driven irrigation architectures. Reported benefits include improved water-use efficiency, enhanced banana yield and fruit quality, reduced input costs, environmental conservation and labour savings. Nonetheless, challenges remain in cost, scalability, connectivity in remote regions, data quality and farmer training. Future research and development should focus on affordable sensor networks, AI-driven predictive tools and capacity building to ensure widespread adoption.

The soil microbial community in coastal wetlands plays a crucial role in biogeochemical cycling. In this study, Spartina alterniflora soil (HB), found near the sea; Spartina alterniflora soil (ZY), found near land; and Phragmites australis soil (KA), found in coastal wetlands, were selected to study the microbial community structure and major environmental influencing factors. The results showed that environmental factors had a significant difference in the three soils. Compared with the ZY and KA sites, the soil at the HB site had the highest value of salinity (14.1 g/kg) and the lowest value of total organic carbon (TOC) (2.9 g/kg) in summer. At the KA site, the values of soil temperature, soil humidity (SH), TOC, and NH4+ were higher than those at HB and ZY sites, while the values of EC (159.8 μS/cm in summer) and salinity (4.4 g/kg) were the lowest. Furthermore, the microbial community structure had significant differences at the three sites. Pseudomonas and Bacteroidota dominated at the HB site, while Chloroflexota and Gemmatimonades were more abundant at the ZY and KA sites. Microbial alpha diversity analysis indicated that the microbial community diversity of Phragmites australis soil was the most uniform, and the microbial species richness in the soil of Spartina alterniflora near the sea was the highest. Salinity, TOC, and SH might be the key environmental factors that affect the structure and diversity of microbial communities in soils. High-salt environments may promote the enrichment of salt-tolerant microbial communities, while high TOC and suitable soil humidity may enhance the uniformity of microbial communities.

The safety of the global food supply depends heavily on effective crop management, making early diagnosis of plant diseases vital for improving agricultural productivity. This proposal outlines the development of an intelligent irrigation system that utilizes machine learning and the Internet of Things (IoT) for the early detection of sugarcane leaf diseases and assessment of their impact on crop yield. The system gathers and analyzes data on soil temperature, humidity, and leaf characteristics-specifically changes in texture and color-using high-resolution photography from unmanned aerial vehicles (UAVs) and IoT-connected sensors. To enhance feature extraction and classification, the system employs a non-linear growing self-organizing map (NG-SOM) embedded within the hidden layers of an artificial neural network (ANN). This advanced model effectively identifies complex patterns in the collected data. Compared to traditional classification methods, this approach achieves a sugarcane disease detection accuracy of 95.6% and reduces false positives by 18.3%. It has been tested on multiple disease types, including red rot, smut, and rust. Additionally, the integration of early diagnosis with intelligent irrigation shows a strong correlation with optimized crop production. Predictive modeling of disease progression based on early detection improves output projections by 22.4%, demonstrating the system's value in precision agriculture. By merging UAV imaging, sensor-based monitoring, and advanced machine learning, this approach offers a promising solution for proactive crop disease management and sustainable yield enhancement in sugarcane farming.

Agriculture is central to Sierra Leone's economy and food security, yet the sector faces persistent challenges including climate variability, soil degradation, and reliance on traditional practices, compounded by limited access to affordable monitoring technologies. This research investigates Internet of Things (IoT) and cloud-based technologies to develop a low-cost, multi-sensor farm monitoring system tailored to Sierra Leone's agricultural environment. The system integrates an ESP32 microcontroller with soil moisture, temperature, humidity, CO₂, and PIR motion sensors for continuous farm monitoring. Sensor data are processed and securely transmitted via Wi-Fi to the ThingSpeak cloud platform using HTTPS, providing time-series storage, real-time visualization, and threshold-based alerts delivered through mobile channels. A design-science engineering methodology guided system development emphasizing affordability, energy efficiency, and usability for resource-constrained environments. Experimental evaluation demonstrated reliable real-time monitoring with sensor accuracy exceeding acceptable thresholds, data transmission latency below three seconds, and stable cloud connectivity under typical rural network conditions. Findings indicate that integrating low-cost IoT hardware with lightweight cloud platforms reduces manual monitoring, improves situational awareness, and supports data-driven decision-making regarding soil conditions, microclimate, air quality, and farm security. The modular design enables adaptation for other developing regions with similar infrastructural constraints. This study demonstrates that an ESP32-based IoT framework with cloud analytics offers a practical, scalable approach to smart agriculture in Sierra Leone, contributing a validated low-cost model bridging advanced technologies with smallholder farmer needs while laying foundations for predictive analytics and large-scale deployment.

Precision agriculture is an enactment of increasing the profitability of crop yields by means of efficient farming practices. In India, farmers can monitor the situations of their surroundings and the ecosystem using precision agriculture in a short period of time. Crop prediction is a critical mission for the decision-makers at the state and district level for speedy decision-making. Therefore, the design and development of an intelligent crop prediction system with high accuracy is a pressing necessity that can assist farmers in determining the crop for cultivation in their fields. In datasets related to farming, different factors like soil nutrients, temperature, humidity, and rainfall commonly depend on each other. A lacuna in the existing crop prediction system is that the correlation between crop features may not be considered, resulting in poor system performance in terms of accuracy. The correlation between features is important as it directly affects the performance of the prediction system. In this study, a Feature Correlation Square based Nearest Neighbor (FCSNN) approach is proposed which extracts the correlation between crop features and predicts the type of crop using the nearest neighbor approach. The proposed crop prediction system is trained and tested using a publicly available benchmark crop recommendation agriculture dataset. It is observed that the proposed approach outperforms the existing crop prediction systems constructed using base classifiers.

Abstract Water shortage is a big problem in farming, especially in states like Maharashtra. Many farmers use too much water because irrigation is not managed properly, which causes water wastage. This paper explains a Smart Irrigation System that uses IoT sensors and Artificial Neural Networks (ANN) to solve this problem. The system uses sensors to measure soil moisture, temperature, humidity, and rainfall. This data is sent to an ANN model, which decides when and how much water the crops really need. The system then automatically turns irrigation on or off. Results show that this smart system can save 35–45% of water compared to traditional methods like fixed timers or simple moisture limits, while still keeping crops healthy. Because ANN can learn from data and improve decisions over time, the system is smart, adaptable, and useful for modern precision farming. Keywords— Smart Irrigation, IoT, Artificial Neural Network, Precision Agriculture, Water Conservation, Automated Irrigation.

Illegal gold mining filters soil bacterial communities and enhances mercury mobility across Brazilian biomes: A multi-season study.

Agriculture remains a key sector of the Indian economy and continues to support a large share of the population for livelihood. However, unpredictable variations in weather and environmental conditions significantly affect crop productivity and yield. Machine Learning (ML) has emerged as an effective decision-support approach for Crop Yield Prediction (CYP), helping improve decisions related to crop selection and cultivation planning based on soil and climatic parameters.Several ML and AIbased techniques have been applied for crop yield estimation, crop classification, and fertilizer recommendation using input variables such as soil nutrients, pH value, temperature, humidity, and rainfall. While Neural Networks show promising results, they often face limitations such as reduced prediction efficiency and difficulty in minimizing prediction error. Similarly, many supervised learning methods struggle to model complex nonlinear relationships between input and output variables. Comparative observations across commonly used techniques highlight the need for more accurate and reliable predictive models to improve agricultural decision-making and sustainable productivity.

Agriculture requires efficient water management and real-time monitoring to improve crop productivity. This paper presents a Smart Agriculture Monitoring and Automated Irrigation System using Internet of Things (IoT) technology. The system uses sensors for soil moisture, temperature, humidity, rainfall, and light intensity, integrated with an ESP32 microcontroller. Sensor data is sent to a local Python server (Flask API) for processing and then stored in a MySQL database. A web application dashboard retrieves data through REST APIs for real-time visualization and remote monitoring. Based on threshold values, the system automatically controls a water pump through a relay module and stops irrigation during rainfall. Manual motor control and alert notifications are also supported. The proposed system reduces water wastage, minimizes human effort, and provides a scalable solution for precision agriculture.

Sustainable agriculture in arid regions faces critical challenges due to water scarcity, high temperatures, and inefficient traditional farming practices. This study presents an AI-enabled smart farming framework for optimizing date palm (Phoenix dactylifera) cultivation through the integration of Machine Learning (ML) and Internet of Things (IoT) technologies. A structured multimodal dataset comprising biometric features palm height, trunk diameter, and leaf number, environmental parameters soil moisture, temperature, and humidity, and categorical attributes variety and health status was analyzed to classify palm health and support data-driven irrigation management. Four ML algorithms Random Forest (RF), Gradient Boosting Machine (GBM), Artificial Neural Network (ANN), and Support Vector Machine (SVM) were developed and optimized using grid search with five-fold cross-validation. Among them, the Random Forest model achieved the highest classification accuracy of 95.3%, demonstrating strong robustness for heterogeneous agricultural data. Feature importance analysis highlighted soil moisture, humidity, trunk diameter, and leaf number as key contributors to palm health prediction. The proposed AI-IoT framework enables real-time monitoring, predictive diagnostics, and automated decision support for sustainable water use and crop management, aligning with Saudi Vision 2030 objectives for technology-driven and resource-efficient agriculture.

Although greening of urban landscapes is an extremely important for the environment formation, less attention is given to it. To improve the urban landscape quality and representation, it is necessary to consider the criteria and characteristics of different plant species when planting them together. This work highlights the consistency of woody-shrubby and herbaceous plants based on such parameters as the root system, illumination, composition, soil humidity and function, winter- and drought resistance, flowering period. Purpose: The purpose of this study is mapping of West Siberian woody-shrubby and herbaceous plant consistency in order to improve the landscapes quality and representation. Methodology : Landscape, cartographic and theoretical analysis, synthesis, comparison, modeling. Research findings: Mapping of West Siberian woody-shrubby and herbaceous plants includes 36 samples and 630 combinations. The compositions are mapped and recommended for joint planting of urban landscapes.

The vine phenological development is closely related to the regional climatic specifics. Changes in temperature, precipitation, and other factors can significantly affect the course of individual phenophases. The red wine varieties Cabernet Dorsa and Cabernet Mitos, characterized by increased resistance to fungal diseases and low winter temperatures, have been selected for specific conditions of countries with a cool climate and a short growing season. The presence of valuable economic and technological qualities in these varieties justifies research related to their regenerative and reproductive performance under conditions of a long growing season and extreme temperatures. The phenological phases and their duration were monitored – bud burst, first leaf appearance, first inflorescence appearance, flowering, veraison, and the onset of technological ripeness. Climatic and soil indicators were recorded - average air temperature, relative atmospheric humidity and precipitation, soil humidity and temperature, while the total and active temperature sum for the onset of the main phases and the entire growing season was determined. The period from bud burst to technological ripeness lasted from 141 to 157 days, which is 30 to 50 days less than the most widely distributed local and introduced varieties. The active temperature sum required to reach technological ripeness is between 1655°C and 1831.1°C. Varieties have low thermal requirements and good adaptation to the climatic conditions of the semi-continental climate. The data would serve to determine suitable terroirs for growing these varieties, depending on the set agrotechnical and technological goals. Keywords: PIWI, phenology, climate change, vegetation, terroir

Water management is a critical challenge in agriculture, particularly for small-scale farms that face resource limitations and unpredictable environmental conditions. Smart irrigation technologies that integrate the Internet of Things (IoT) and machine learning offer significant solutions in enhancing water efficiency and boosting crop production. This study investigates the synergistic application of IoT-enabled sensors alongside machine learning methodologies, specifically Decision Trees (DT) and Support Vector Machines (SVM), to augment irrigation effectiveness. Real-time sensor data collection, featuring elements like soil moisture, temperature, and humidity, serves to direct irrigation techniques. The proposed utilizes solution supervised learning techniques to establish optimal irrigation timetable and reinforcement learning to modify decisions based on real-world performance. Preliminary findings suggest that SVM outperforms DT in reducing false positives and negatives, leading to more precise irrigation control. The study underlines the benefits of AI-driven irrigation system, such as enhanced water conservation, higher crop yields, and increased sustainability. Furthermore, the difficulties of establishing IoT-based irrigation systems, such as data security, connectivity constraints, and cost considerations, are addressed. The findings add to the literature of precision agriculture and provide useful insights for small-scale farmers who are willing to implement smart irrigation solutions. The study's goal is to enhance efficient water use, strengthen food security, and support sustainable farming methods by combining IoT and AI. To get the most out of AI-powered irrigation systems, future research should focus on enhancing algorithm accuracy, expanding real-world trials, and tackling scalability challenges.

IoT Based Smart Crop Field Monitoring & Automation of Irrigation System offer a trans formative approach to optimizing agricultural practices, allowing farmers to achieve enhanced precision, efficiency, and sustainability. The system incorporates a network of sensors deployed throughout the crop field to collect real-time data on crucial parameters such as soil moisture, temperature, humidity, light intensity, and nutrient levels. The data is transmitted wireless to a central control unit, where it is processed and analyzed using advanced algorithms. Remote monitoring and control capabilities enable farmers to access this data and manage the irrigation system from anywhere using their smart phones, tablets, or computers. Automated irrigation control is a fundamental feature of the system, leveraging collected sensor data and predefined parameters to determine the optimal timing, duration, and quantity of water required by the crops. Weather integration plays a crucial role in the system’s operations, incorporating real-time weather data and forecasts. Data analytic capabilities enable farmers to gain valuable insights and trends regarding crop health, water usage, and growth patterns. The system emphasizes energy efficiency by intelligently managing irrigation operations and can integrate with energy management systems and employ energy-efficient components to schedule irrigation cycles during off-peak hours and conserve energy consumption. The systems revolutionize traditional agri- cultural practices by providing precise monitoring, automated irrigation control, weather integration, data analytic, energy efficiency, scalability, and flexibility, allowing farmers to optimize irrigation management, enhance crop yield, conserve resources, and contribute to sustainable agriculture.

Climate change consequences assessing for environment and humans is one of the main tasks facing ecology and has great practical significance. Long-term studies of small rodent populations have contributed fundamentally to the development of population ecology. Extreme drought in the forest zone of the Southern Urals in 1975 is one of the rarest phenomena, a climatic event (once in a century), and it is considered as a contrasting background that provides for especially clear manifestation mechanisms of adaptive resistance in animals populations. In the paper, we present for the first time the 52-year dynamics of species structure of the rodent community and population abundance data, obtained in the Ilmen State Nature Reserve (Chelyabinsk Region, Southern Urals, Russia) before and after the drought of 1975 and its remote consequences. We tested the hypothesis about the similar population responses to both an extreme drought and normal autumn-winter conditions. The drought was verified by a dendrochronological indicator (radial growth of Pinus sylvestris trunks) and two climatic parameters, namely soil humidity (for the first time) and precipitation amount. Population dynamic phase portraits demonstrate species-specific features concerning mainly abundance levels before and after the extreme episode: dynamic stability of Clethrionomys glareolus and C. rutilus; abundance increase of Sylvaemus uralensis; collapse of population abundance of Microtus agrestis, M. arvalis, and Alexandromys oeconomus. Species composition (local species lists) turned out to be a reliable indicator of drought. Based on both original functional-ontogenetic approach and individual marking (CMR-method) data, the species populations' response patterns to drought were studied. In summer 1975, the Clethrionomys species have implemented a response strategy, which historically developed for regular winter conditions, reflected in complete blocking of yearlings sexual maturation (minimisation of metabolic processes) as we hypothesised originally. The next year, we observed population adaptations included breeding period prolongation of overwintering females and intergenerational crossing (age cross), which provide effective utilisation of possibilities for population growth along with preservation of yearlings and improvement of genetic heterogeneity. The Sylvaemus uralensis population demonstrated a similar response pattern to drought, reflected in partial blocking of yearlings' sexual maturation, then a gradual increase in the number of individuals up to a high dynamic level in recent decades. In contrast, Microtus agrestis, M. arvalis, and Alexandromys oeconomus populations are characterised by the usual participation in reproduction of both overwintering individuals and yearlings in drought year (i.e. our hypothesis has not been confirmed). As a consequence, there was a mass mortality and abrupt transition of the populations to the critically low abundance level of its oscillations and elimination. An invasion of an alien species (Apodemus flavicollis) was recorded for the first time in 2020. The weather anomaly triggered rapid long-term changes in the biotic community structure, by reflecting in the increase in population abundance of granivorous/seed-eating species (Clethrionomys glareolus, C. rutilus, Sylvaemus uralensis), the finding of a new species (Apodemus flavicollis), and changing of co-dominant (Sylvaemus uralensis instead of previous Microtus agrestis), and, finally, disappearance of herbivorous species (Microtus agrestis, M. arvalis, and Alexandromys oeconomus). Thus, long-term monitoring of rodent populations in forest ecosystem of the Ilmen State Nature Reserve revealed a real possibility of the rapid population rearrangements on evolutionary scale in populations of the local fauna as a result of extreme drought. Drought exposure consequences in the studied Rodentia species are markedly extended in time, being fixed in a series of generations.

ABSTRACT Recently, precision agriculture has used wireless sensor networks (WSNs) to gain valuable insights and improve crop yields, promoting efficient resource use and data‐driven decisions. However, WSNs face challenges, such as high power consumption from continuous sensing, data processing and communication, especially in large‐scale setups, which limits their lifespan. This paper focuses on reducing power use in agricultural WSN sensor nodes during data transmission of soil moisture, rainfall, light intensity, air temperature and humidity from the transmitting sensor node to the base station. Four algorithms are proposed to cut power consumption. First, a sleep/wake (S/W) scheme using a simple duty cycle called S/W‐DC. Second, the S/W scheme combined with adaptive data sampling (ADS) based on redundant data (RD), called S/W‐ADS‐RD. Third, the S/W scheme integrated with dynamic voltage scaling (DVS), named S/W‐DVS. Fourth, a hybrid of all three, called S/W‐ADS‐RD‐DVS. The sensor uses a 12 V/5 W solar panel for energy harvesting to maintain operation. The hybrid algorithm achieved 99.232% power savings and extended battery life to approximately 1.83 years. During a 6‐h session, data transmission was reduced by 99.93%. This research could significantly improve WSN efficiency in precision agriculture and can be applied to energy‐efficient WSN deployment across various fields, supporting Internet of Things (IoT) applications.

Precision irrigation is essential for sustainable agriculture under increasing water scarcity. This study compared regression and ensemble learning algorithms for forecasting irrigation requirements in sweet potato, a crop characterized by high variability in water demand. An Internet of Things (IoT)-based prototype was deployed to collect real-time data on soil moisture, temperature, humidity, light intensity, and atmospheric pressure over 42 hours and 50 minutes (August 4-5, 2025), encompassing two complete diurnal cycles at 10-minute intervals and yielding 243 temporal observations. Following preprocessing and feature engineering with lag-based temporal features, the final dataset comprised 240 samples (192 training, 48 testing) using chronological time-based splitting to prevent data leakage. Five algorithms, Support Vector Regression (SVR), AdaBoost, Extreme Gradient Boosting (XGBoost), Random Forest Regressor (RFR), and CatBoost, were evaluated under default and hyperparameter-tuned configurations using Root Mean Squared Error (RMSE), Mean Absolute Error (MAE), and Coefficient of Determination (R²) as evaluation metrics. Tuned Random Forest achieved superior performance (R² = 0.9802, RMSE = 9.58, MAE = 6.08), followed by default Random Forest (R² = 0.9786) and default CatBoost (R² = 0.9687). XGBoost demonstrated strong performance (R² = 0.9670 tuned) but exhibited overfitting tendencies with near-perfect training scores. SVR improved substantially after tuning (R² = 0.328 to 0.797), although it remained inferior to ensemble methods. Overall, ensemble methods, particularly XGBoost and Random Forest, demonstrated superior efficacy for sweet potato irrigation forecasting. These findings underscore the potential of IoT-integrated machine learning to enhance water-use efficiency and support sustainable smart farming practices.