Digital frontiers in agricultural and biological innovation

Introduction

Agriculture is shaped by balance between nature's biological processes, the limits of environment, and the socio-economic pressures making them constantly changing result [1,2] . The agricultural and biological systems are basically complex, dynamic, and multiscale, shaped by functioning among environmental factors, genetic variability, management practices, and socio-economic influences. The global demand for the ethical, sustainable and resilient food systems is increasing strongly day by day due to challenges and problems such as environmental degradation, growing population and increasing issues in public health [3] [4] . The recent studies highlight the importance of integration of new advanced computational techniques with specialized knowledge to address and overcome these challenges or limitations to create new opportunities for unlocking the sustainable innovations of Agriculture [5,6] . The modern agricultural modelling has evolved from empirical and static framework towards the data intensive, process-oriented systems that uses the extensive datasets obtained from remote sensing, modern sensors, and environmental monitoring to intensify prediction accuracy and understanding systems across various scales. New Innovations in Precision agriculture have revealed that deep learning and artificial intelligence techniques can help in improving crop monitoring, yield forecasting, and anomaly detection by capturing nonlinear relationships that includes traditional statistical models. Currently at the same time, the hybrid methods approach that integrate mechanistic process models with machine learning have demonstrated more robustness and generalizability, facilitating scalable and interpretable decision support systems. Optimization techniques are also very advancing. The Multi-objective structures that balances productivity with sustainability indicators such as consumption of water, greenhouse gas emissions are becoming essential tools for making resilient agricultural systems in face of climate change. The resource allocation and crop planning models utilize optimization algorithms offers enhanced strategies for climate adaptation and utilization resources efficiently. As complexity for modelling increases, future studies must prioritize the integrating AI-assisted modelling with optimization to facilitate real-time, adaptive, and scalable solutions for sustainable agriculture. Combining the AI, all the big data and biological informed models open inspiring path for both agricultural and biological systems it helps in creating more sustainable, strong and smarter solutions so the population and the planet can rely on.

Evolution of Biological and Agricultural Modelling

The steady evolution of the fields of biological and agricultural modelling for over a century mirrors the rising complexity of living systems and also increases collaboration between scientific disciplines. The initial biological models, based on mathematical biology, designed to comprehend population growth dynamics and physiological processes, offering mechanistic insights into systems of nature. The origins of agricultural modelling can be traced back to empirical methods designed to solve practical problems that farmers faced when managing crops and farms. It evolved into progressively more intricate process based and system models overtime, taking into account management variable, soil conditions, and climate [7] [8] .Continues advancements in computation and system research have enable the development of integrated, multiscale modelling frameworks that support sustainability assessments, planning for climate adaptation, and decisionmaking within modern agricultural systems [2] .

Early Conceptual and Mathematical Modelling in Biology

Early conceptual and mathematical modelling in biology emerged from the quest to better understand the behaviour and evolution of living systems. In the early to mid-20 th century, scientists started examining processes like disease transmission species relationships, population growth, and physiological control using mathematical methods like differential equations and dynamic systems theory [9] . These techniques allowed scientists to go beyond simple observations and characterize biological change in a predictable and quantitative way. One of the early mathematical biology to represent feedback loops, stability, and nonlinear behaviour all characteristics of actual biological systems was one of its main advantages. Traditional population models helped scientists make sense of patterns seen in nature, they described why populations cannot expand indefinitely and how competition or predator-prey interactions can lead to cycles and oscillations. These models represented a significant change in biology toward analytical reasoning and hypothesis testing, despite the fact that they were occasionally oversimplified and based on ideal conditions. Early models provided a solid conceptual basis for subsequent advancements in ecological modelling, systems biology, and process-based simulations by converting biological processes into mathematical form. Ultimately this advancement enabled more integrated and realistic modelling techniques in the modern biological and agricultural sciences by making it possible to connect processes at various levels of biological structure [10] .

Emergence of Process-Based Crop and Agricultural Models

When process-based crop and agricultural models emerged, agricultural systems analysis underwent a fundamental shift from purely empirical descriptions to mechanistic representations of the biological and physical processes governing crop development and output. The majority of the early crop models created in1960s and 1970s, according to Sinclair and Seligman (1996), were site-specific and statistical, based on past yield-weather connections, had no ability to extrapolate beyond observed conditions [11] . However, growing climate variability, resource limits, and the requirement for predictive decision-support systems prompted models based on physiological and ecological principles.

In order to respond dynamically to environmental and management inputs, process-based models explicitly replicate important processes as photosynthesis, respiration, phenological development, biomass partitioning, soil water balance, and nutrient cycling [12,13] . The theoretical foundation for connecting radiation interception and biomass accumulation was developed by the seminal work of de Wilt (1965) and later Monteith (1977), and it proved essential to many crop development models. These ideas made it possible for models to represent cause and effect links in crop-soil-atmosphere systems rather than just correlations. With the introduction of major recreation platforms such as DSSAT, APSIM, and WOFOST, which connected crop physiology with soil and weather modules to evaluate genotype-environment-management interactions in multiple agroecological zones, process-based modelling goals grew dramatically in the 1980s and 1990s. Similarly, APSIM integrated crop rotations, soil processes, and management decisions to support sustainable land-use planning by emphasizing the changing aspects of the whole farming systems [7,8] . These models in particular showed improved predictive accuracy when evaluating yield response to nitrogen and water availability under different management and climate circumstances [2] .

Systems and Integrated Modelling Approaches

As scientific understanding of agroecosystems expanded, agricultural modelling evolved from isolated, componentspecific assessments to systems and integrated modelling methodologies. Agriculture works as an important living interconnected system rather than a set of separate components [2] . All crops, soils, animals, climate, pests, management practices, and socio-economic factors continuously interact with one another, changes in one element affect the others. Integrated models capture this complexity by viewing agricultural production is not just biophysical process but also a human-natural system in which ecological processes are influenced by farmer decisions, market forces and policy interventions. Approaches such as Integrated Farming System, agroecosystem, and whole farm models evaluate environmental impacts, productivity, and efficacy simultaneously [14] . This perspective of understanding sustainability is essential, identifying trade-offs, and strengthening long-term system resilience of farming systems.

Data-Driven and Statistical Modelling

The statistics and data-driven models began to play a considerably big role in agricultural analysis as computer became faster and data became more plentiful. Now, Researchers could use many tools such as regression analysis, stochastic methods and time-series models to make sense of complex patterns in crop yield, weather variability, and production risks [11] . Deep biological knowledge of crops or soils were not required for these models as they were appealing and they could be implemented quickly across larger regions using available datasheets. They were especially useful for risk assessment, large-scale planning and policy analysis. However, this strength was also a downside though. As these models heavily depends on past data and they also think that future conditions will be similar like past. At the time that agriculture faces new challenges and problems like change in climate, extreme weather events or fast change in farming practises their predictions can become uncertain [3,11] . Consequently, while data-driven models remain valuable tools, because they are most effective when complemented by process-based and systems-oriented methods that can better handle in unaware conditions [2] .

Integration of Remote Sensing, GIS, and Big Data

The introduction of sensors, GIS, and remote sensing technologies revolutionized the way of management and understanding of agricultural systems. This allowed farmers and researchers to know about conditions happening across entire fields, regions, and even countries, instead of relying on field observations or average values [1] [15] . Because of satellite images, drone-based data, and real-time sensors it made much more possible and precision to track crop growth, soil moisture, and land-use changes with great detail and accuracy [5] . This rapid change brought agriculture closer to real-world circumstances due to the fact that no two fields behave precisely the same. There are early detections of variations in soil type, water availability, and crop health enabled more informed and more timely decision making. Precision agriculture models, came out of this technological progress which emphasize the field management site by site rather than treating them uniformly. With the help of combining spatial data with statistical and simulations models farmers now can decide where to irrigate, fertilize, or protect crops in order to reduce waste and increase efficiency. [15] . While researchers at the same time gained powerful tools and instruments to study agricultural systems on a greater scale. Overall, because of these developments both the farm-level decision making and regional to global evaluations, making agricultural modelling more useful, scalable, and closely related to actual farming environments.

Artificial Intelligence and Hybrid Modelling Frameworks

Recently the advances in deep learning, machine learning, and artificial intelligence have brought up a new level of biological and agricultural modelling intelligence. The researchers and farmers can predict yields, detect early diseases attack, and crops response to climate change conditions because of these tools for large and complex datasets are especially good at finding hidden patterns [5,6] . The hybrid models deliver more accurate predictions, after connecting the strengths of both approaches, while respecting biological principles. The real-world decision making has valuable balance between accuracy and understanding which makes them special [16] . In result, hybrid modelling framework plays a major role in bringing the agricultural and biological modelling closer to not only smarter systems but also to systems that are transparent and scientifically meaningful.

Current Trends and Future Directions

In today's time the biological and agricultural modelling is now moving through beyond prediction and also into the realm of decision-making practically. Sustainability, climate resilience, and long-term system health are main focus now [2] . New models and tools are now used to help farmers for their crops, nutrient and water management more efficiently with reduced environmental impacts and adapt these practices to increase climate uncertainty. While policymakers also depend upon these tools to explore the consequences of different strategies before implementing them on the main field. The new models, tools and instruments of technology are designed to support the everyday decisions and provide solution to the challenges and problems faced by farmers. These future models are becoming more adaptive and broad day by day. The sensors and satellites provide the real-time data which enable models for continuous update, while it brings farmers and other stakeholders as participatory approaches in modelling processes. The aim of socioeconomic and policy dimensions remains combined to make more transparent, accessible, and actionable models, ensuing the delivery of meaningful benefits for farmers, researchers, and decision-makers [16] .

AI-Powered Hybrid Modelling approaches in Biological and Agricultural Systems

Since AI-driven hybrid modelling framework in biological and agricultural sciences are gaining attention by combining the two best different ways of complex systems understanding. On one side models based on wellestablished biological, physical, and ecological principles concepts are referred to as mechanistic and process-based models [7] [17] . These valuable Models helps in explaining the system behaviour and way of work, offering transparency, interpretability, and casual understanding. However, such models can falter in actual agricultural environment settings where change in climate, management diversity, and biological interactions is rapidly shifting [18] . There are challenges to rely on parameters like measure them accurately and it may oversimplify processes that are not fully understood. Approaches of data-driven which includes machine learning and artificial intelligence are extremely good at identifying patterns and making predictions for large datasets [19,20] . By allowing data-driven components, hybrid models bridge gaps to complement process-based knowledge. In recent studies suggests that such integration is especially effective where models must remain flexible yet biologically realistic under climate uncertainty and in data-rich environments [16,21,22] . As for the decision support, sustainability evaluation, and the designed resilient agricultural systems it is seen that hybrid frameworks is increasing practically and trustworthy foundation for them [23] .

Hybrid Integration of Mechanistic and Machine Learning Models

Today there is practically better way to understand and predict complex agricultural systems with the help and utilization of hybrid integration of mechanistic and machine learning models. These mechanistic models provide description about crops and biological systems functions using scientific principles, such as crop development stages, photosynthesis process, soil water movement, and nutrient cycle [17,24] . Biological explanations are strongly provided by these models, climate change, spatial differences, and incomplete data are caused so they often struggle to represent real-world variability [18] . Through large datasets learning patterns are complemented by machine learning models, due to lack of biological reasoning [19,20] . The machine learning algorithms of data with biological knowledge both approaches are strength of hybrid models.

For improving the prediction accuracy and stability particularly for crop yield estimation this integration has been implemented. Hybrid frameworks improve performance by reducing overfitting in unfamiliar conditions and better embeds mechanistic outputs or constraints into machine learning models. In data-limited or climate-stressed regions these models are very valuable, where AI learning understood scientifically. As a result, integration of hybrid models provides a balanced approach that improves reliability while maintaining scientific credibility.

Bi-Stage and Multi-Component Hybrid Frameworks

In an organized and explainable way bi-stage and multicomponent hybrid frameworks form the interaction between AI and mechanistic models. The mechanistic models are used first in order to simulate meaningful biological indicators like biomass growth, stress levels, or crop phenology in bi-stage systems. For machine learning models the outputs are used as inputs for refining further predictions by capturing complex interactions. AI is allowed to focus on improvement of accuracy without disregarding biological meaning of this approach. Several different and unique models such as climate models, crop growth simulators, and AI based prediction models are linked together into single approach to expand its ideas for multi-component frameworks [25] . Hybrid frameworks are mainly utilized in plant breeding for genotype prediction and environmental interactions which is bit problematic and field experiments are really expensive [26] . Hybrid models fasten up the breeding decisions and also improved prediction consistency by dropping dependency on large scale trials across environments. Their models design allows individual components now updated by new data and improved algorithms are also available which makes them more flexible. In conclusion, these hybrid frameworks provide flexible solutions for complex biological systems.

Hybrid Methods for Yield and Quality Prediction

The models of hybrid frameworks for yield and quality prediction is used not only for prediction of crop yield but also concerns about quality traits like protein and micronutrient content in the crop. Before yield quantity was mainly focused by the agricultural models traditionally dominating the quality parameters which are required for nutrition, market value and food security [27] . Among genetics, environmental conditions, and management practices quality traits are mainly influenced by complex interactions which makes problem for them to make model accurately. Further, for this problem hybrid frameworks provided solution that by combining process-based models which include crop growth simulations and nutrient uptake through machine learning models captures non-linear relationships from field observations and remote sensing data [28] . By integrating this approach, it enables more detailed surveillance and spatially explicit predictions ensuring both yield protection and high-quality standards. By aligning with physiological processes, studies demonstrates that hybrid models framework provide accurate and improve predictions of micronutrients such as iron and zinc. By utilization of satellite and sensor data provides critical infrastructure required to support largescale applications. These new advances in frameworks support the idea of precision nutrition, biofortification programs, and better crop management decisions [29] . Consequently, hybrid models for yield and quality production transforms agricultural assessment to a holistic evaluation of food quantity and nutritional value from productivity-based metrics.

Explainable AI in Hybrid Frameworks

It is very essential for building trust in AI-driven hybrid models for explanation very important while support decisions of these models are made by farmers, researchers, and policymakers. Some results are difficult to interpret because of complex AI components while on the other hand prediction accuracy is improved with the help of hybrid models. This explainable AI in hybrid frameworks also provides solution to address the problem while clarifying that how modular outputs are influenced by different input variables [30] . For finding the influential factors which affects yield, stress response, and disease risk mostly some hybrid frameworks tools are used increasingly which are SHAP and LIME. This view helps the users to understand the particular predictions produced by the model. Within known biological process rather than relying on misleading corrections this explainability hybrid framework supports the scientific validation and also by aligning the AI-driven insights. It is particularly important in agriculture because sometimes decisions involve risk, cost and environmental impact. With responsible use of AI models and wider adoption encouragement through hybrid models these predictions are more understandable and explainable with the help of explainability AI frameworks. Better improvement of communication in between model developers, decision-makers and users because of combining explainability which strengthens confidence in hybrid systems.

Advanced AI Architectures and Hybrid Deep Learning

In hybrid model frameworks, Advanced AI architectures and hybrid deep learning models are becoming central and increasingly used to understand the complex biological and agricultural systems dynamics. Deep learning model is ideal for managing the vast datasets on large-scale and also highly effective for modelling non-linear relationships with time sensitive dependencies interactions between climate-soilcrop growth. Biological structure and data-driven flexibility both are benefitting the hybrid deep learning systems if combined with mechanistic models [31] . By the use of these high technology models many improvements are in the performance of yield prediction, stress detection, and climate impact analysis under extreme and unfamiliar conditions. Deep learning improves the model adaptability with the help of identifying residual patterns and refining mechanistic outputs, while ensuring the preservation of scientific constraints. Large amounts of datasets, significant computing power, and careful calibration is required and utilized by these models. The latest studies focused on transfer leaning and physics-informed neural networks and advanced training methods to reduce the high problems with less complexity.

Hybrid Predictive Systems in Precision Agriculture

In hybrid predictive systems, by integrating AI, mechanistic models, and real time data precision agriculture is transformed into practical decision-support. These hybrid systems in precision agriculture collects the combined information with the help of weather stations, soil sensors, satellites, and farm records which is further used to provide specific location detail [32] . The newly updated data becomes available continuously through AI algorithms and crop water relation, nutrient content, and climate conditions explained by mechanistic models. For betterment of irrigation, fertilization, and disease management IoT devices and edge computing is used to allow this system to operate on time. These hybrid predictive systems help the farmers to optimize inputs and effective at managing field-level variability. These frameworks convert the model output into actionable guidance and also work like a gap between realworld farming and research. Bridging these gaps will be important to overcome these barriers and obstacles which is the key to maximize the profit and impact of precision agriculture in hybrid systems. As the utilization of advanced technology rapidly increasing day by day researchers introduced climate-resilient agricultural modelling which helps the farmers and provides better understanding of farming systems affected by climate change and ways of sustainability adaptation in agriculture [33] . These models were essential and much needed for the climate variability, frequent extreme events, and growing pressure on natural resources also for sustaining future and supports practical decision-making [34] [35] . To surge predictive accuracy and resilience, future research is merging advanced computational tools with traditional and ecological process models by researchers [33] . The integration of process-based crop models and machine learning is a key advancement in the field of technology. These process-based models are rooted in plant physiology and ecological processes, enabling them to predict crop growth simulations and yield responses to variables like temperature, water availability, and carbon dioxide levels [36] . Extensive data and complex calibration requirement limit the application in data-scarce regions of these models having strong theoretical knowledge. Hybrid models bridge the gap to address these limitations by merging mechanistic knowledge with deep learning [37] . Machine learning reveals complex climate-crop predictions with non-linear relationships unlike traditional models, ensuring biological realism [5,37] . That is the reason of hybrid models perform better than other approaches especially climate-resilient modelling in agriculture.

Innovations in Climate-Ready Agricultural models

The increasing application of explainable AI and artificial intelligence assess crop performance under climate stress is closely related to development [38] . AI-driven models depend upon Remote sensing, satellite imagery, and historical climate data allows to predict [38,39] yields, soil and climate conditions, and pest management. AI helps farmers, advisors, and policy-makers by gaining trust and clarifies about certain predictions of models. With the help of these models and tools it is easy to translate the complex data into understandable insights which supports management and sustainability of precision agriculture. By merging the data from different sources with the help of data fusion and remote sensing is another future insight of development. The more accurate data is provided by monitoring of crop phenology and detection of stress earlier with the help of combined high-resolution satellite data and ground-based sensors. By incorporating this data into crop model, it allows to detect droughts, heat waves, and moisture loss early prevents major yield loss. Such early warnings are necessary for risk management and losses [40] .

In future research multiscale and multi-agent modelling approaches will be prioritize more. By overcoming the traditional models which operates at single spatial scale, emerging frameworks begins to link farm-level processes with regional and national food systems. For climate shocks, researchers introduced agent-based models to characterize farmer decision-making, technology adoption, and socioeconomic responses [41] . These models provide more real image of resilience results and policy effects while merged with bio-physical models [42] .

Researchers are now gaining attention by combining genotype-environment-management modelling which connects crop breeding with climate adaptation [43,44] . By modifying genetic traits, management practices, and environmental variability composed, these models further helps by identifying characteristics of different crops for improving the heat and drought tolerance, guiding future breeding efforts [44] . As a result, open science and collaborative technology are recognized as vital for progress [33] . By promoting shared datasets, interoperable platforms, and community-driven development improves the transparency, reproducibility, and global accessibility, ensures Climate-resilient agricultural models research impacts policy and practice [45] .

Smart Optimization for Green technologies

As agricultural system is growing more complex day by day, multi-objective optimization is becoming more essential for management of complex sustainable systems [36, 46] . Today's agriculture prioritizes a balance between climate resilience, resources efficiency, and socio-ecological stability, it is no longer focuses on yield maximizing. To address the pressure from growing climate conditions, environmental degradation, and socio-economic constraints there is a certain requirement of modern approaches to evolve with time [40] . Unlike single-objective approaches, multi-objective optimization helps balance competing goals like profitability, environmental protection, and social wellbeing making it more effective [46] . By enabling a better understanding of real-world trade-offs, this approach helps researchers and decision makers to design long-term productivity, resilient, and sustainable agricultural systems [36] .

Advanced Integration of Biological Process Models and Optimization

The future direction in biological and agricultural modelling is all about the combining of process-based models with multi-objective optimization frameworks [36] . Before, biological models were used for describing crop growth, soil carbon dynamics, nutrient cycling, or pest-disease interactions for simulation and optimization applied as separate analysis [47] . Modern approaches integrate optimization with biological models, and eliminates the traditional separation [36] . This enables the management of irrigation scheduling, fertilizer application, and varietal selection based on system feedback [48] . This advanced integration of biological models better reflects the realworld decision-making and helps farmers and researchers to respond changing environmental and biological conditions rather than making assumptions. These modern approaches also enhance prediction accuracy for the assessment of longterm sustainability and takes an important step toward advanced system-oriented modelling approaches.

Multi-Scale and Multi-System Optimization

In agricultural system, the sustainability challenges operate across multiple-spatial and temporal scales, which ranges from physiological plant-level processes to landscape and regional food system changes. With the help of multi-scale and multi-system optimization future research is designed to address this complexity [36,40] . While field-scale optimization focuses on targeted crop management and on the other side farm or watershed scale includes land-use allocation, livestock integration, or water sharing among users. Regional planning objectives includes food security, ecosystem service provision, and resource conservation at border scales. These scales are combined under an optimization framework which further allows assessment of trade-offs and avoid local solutions that performs poor at higher levels. These frameworks are vital for maintaining the longevity of soil health, water availability, and biodiversity which is very helpful for maintaining sustainability.

Incorporation of Climate Change and Extremes

As climate changes is creating unpredictability into biological and agricultural systems, making traditional production optimization strategies stability for future is no longer a viable approach [45] . The future multi-objective framework is designed to specifically account for extreme events and climate variability such as droughts, heat waves, floods, and pest outbreaks [40] . Rather than focusing on normal conditions, these approaches focus on robustness and resilience and they look for solution that works well in future. This approach requires combining climate projections, stress-response relationships, and risk-sensitive objectives into models. This approach allows decisionmakers to evaluate strategies based on expected outcomes and their ability of resilience to stress. Such optimization is relevant for climate-vulnerable regions, where risk reduction and adaptive capacity is as important as productivity.

Precision Agriculture and Spatial Modelling

The integration Precision agriculture and spatial modelling signify farmers are moving toward more precise and uniform field management, site-specific decision making [43] . In this approach the use of inherent variety present in soils, crops, and microclimates to inform management decisions. Data is collected about spatial and temporal information with the help of advanced remote sensing, geographic information systems (GIS), global navigation satellite systems (GNSS), and proximal sensing to maintain and get information on crop health, soil properties, moisture conditions, and nutrient dynamics [39] . Spatial modelling is about combining geological referenced datasets by using statistical, geostatistical, and processbased approaches to get better details and understanding about crop growth and yield of different fields [37] . These models represent an idea of sustainability and resource-use efficiency and supports proper use of inputs with proper rate of application such as water, fertilizers, and pesticides to reducing environmental losses. With the help of machine learning technology spatial modelling strengthened to a great aspect as it includes identification of hidden patterns, prediction of yield gaps and detection of stress zone at early growth stages of crop which is very helpful and vital for agricultural systems. The precision agriculture and spatial modelling by integrating contributes to achieve more sustainability in future also and includes agricultural management [37,38] .

Spatial Variability Characterization and Field Zoning

Every Agricultural field is not same, they have variability in soil type, moisture in small plots [49] . Precision agriculture helps in identifying about the difference occurrence in the field with the use of modern technology. Based on yield potential, nutrient status, and water availability practical management zones are made with the help of spatial modelling [50] . Instead of treating all entire field similarly, farmers now can adjust inputs to match their required needs [51] . By allowing this approach it improves efficiency, reduces input wastage, and supports more balanced crop growth across the field.

Geospatial Data Sources and Integration

Precision agriculture today relies on spatial data sources and combines multispectral data, images from satellite, local weather data, drone imagery and yield mapping [52] . Each and every source has its own way of capturing and collecting data or information from different aspect of field conditions. By allowing datasets together with the help of spatial modelling the use of GIS-based platforms, the patterns can be viewed and analysed in a unified way [53] . The integration of different data sources advances in management decisions, soil characteristics for long-term and short-term crop responses are also considered. Merged effective data turns complex spatial information into understandable insights useful for farmers [54] .

Crop Growth Monitoring and Yield Prediction

Crop growth and yield prediction is easily monitored and observed with the help of spatial modelling approach, also observes the crop response to stress across a field [15] . For tracking and managing crop vigour and identifying early signs of nutrient deficiency, water stress, and diseases remote sensing indicators are used [55] . Spatial models can estimate accurate data when soil and weather data is combined [11] . These models provide predictions which helps farmer at proper time which further improves crop performance and reduces yield losses [33] .

Variable Rate Technology and Decision Support

Variable rate technology applies spatial modelling directly on farm actions. Spatial analysis maps drive variable amount of application of inputs like seeds, water, fertilizers, and pesticides [56] . By considering both agronomic and economic constraints, these tools help to turn map data into practical decisions. Studies indicates that precision agriculture responds to within field differences rather than using blanket applications that may be risky to use and also helps in reducing environmental losses, lowers input costs while sustaining high yields [40] .

Environmental Sustainability and Risk Management

Using spatial models helps agriculture improve environment sustainability and reduce risks. Identification of areas is done which are prone to erosion, nutrient runoff, and water stress, so farmers can easily focus on conservation measures [57] . By managing reduction of negative impacts on soil and water resources. Spatial techniques are important for managing climate-driven hazards like drought, heat and flooding [58] . Precision agriculture increases productivity growth which enhances environmental sustainability, and makes farming systems more resilient [59] .

Digital Agriculture and Rural development Modelling

Integrated farming systems creates a sustainable holistic approach by combining different enterprises of agriculture such as crops with livestock and trees, poultry, and forestry by keeping in mind the idea of recycling [14] . Integrated farming also helps in reshaping farm-level socio-economic dynamics, making tools for monitoring sustainability and scalability with the use of socio-economic modelling [60] . It includes the influence of integrated systems on farm income stability, labour use, risk distribution, and household resilience in agricultural systems [61] . Integrated farming systems are useful as it generates employment and multiple income streams in whole year and takes important decisions for farmers welfare. To change the complex data into simplest form socio-economic modelling is used, and also includes farm household models, system dynamics, and agent-based models. By simulating farmers decisions under resource constraints such as land size, capital, labour, and market access and capturing for the feedback between production, investment, and consumption. As a result, these insights highlight the role of integrated farming system in increasing livelihood security, generating employment for women and youth and efficient use of family labour [61,62] . At broader scale, socio-economic modelling helps supporting policy analysis by observing trade-offs between productivity, equity, and environmental decisions. These models also analyse how incentives for livestock integration influence farm profitability and regional resource use. The integration of socio-economic approach with biophysical models enables more accurate assessment of integrated farming system.

Uncertainty, Risk, and Ethical Considerations

Uncertainty, risk and ethical considerations are naturally occurring parts of farming systems. Resources used in agriculture is not fully predictable because they work differently, so farmers use management practices to maintain balance between each and every aspect. Models also carry uncertainty from many different sources such as imperfect data, incorrect assumptions, problems in parameters estimation, and no idea about climate conditions change. It is very important to keep these things in mind otherwise model results may be not good or misleading [2] . Researchers now use probabilistic, stochastic, and ensemble approaches to predict different outcomes instead of giving one exact prediction and this step helps users to get better understanding of different situations that can occur in future [63] . For extreme events including drought, heatwaves, flood, and fires risk-based models is there for climate change conditions they occur rarely but have serious impact on food security and livelihood of farmers. As we know agricultural models and technology are used increasingly in policy design and resource allocation. This makes it vital that model results are reliable, transparent, and responsibly understood. The ethical concerns are also important, small farmers and vulnerable communities are affected by models, challenges of fairness, inclusiveness, and accountability become serious [64] . New ethical challenges also introduced by Artificial intelligence, models based on AI sometimes may shows data bias, lacking clear explanations, and accessible to only those having digital resources which increases inequality. But ethical frameworks also encourage stakeholders, explain uncertainties, and utilization of model outputs carefully. The responsible approaches build trustworthy relationship [65] . Uncertainty, ethics, and risks addressed together helps and supports in ensuring agricultural modelling fairly and decisions responsible socially.

Future Research Gaps and Challenges 8.1 Data limitations in real farming conditions

Some of the agricultural models still depends upon experimental data, and real farms functions under highly flexible conditions [66] . Long-term and high-quality datasets are uncommon in some regions and countries which under development [67] . The important information on soils, weather conditions, farming system, and socio-economic conditions sometimes remain incomplete.

Difficulty in linking scales-Spatial and Temporal

Agricultural processes operate at every scale from plant physiology to global food systems. The shift in advancement of biological and agricultural is bit challenging also they face some unresolved issues. Sometimes models often fails to bridge the gap between spatial and temporal scales, and remaining models restricted to either local plot processes or broad regional assessments. These scales are essential to connect farm-level decisions with landscape, national, and global sustainability ideas. Collecting data together from satellites, field observations, and farmer surveys also seen as a major challenge in future modelling.

Limited inclusion of human behaviour

Optimal recommendations are not always applied in practical farming and not followed by farmers every time. Their choices depend upon experience, risks observation, different resources, and social factors [68] . Realistic modelling requires better and enhanced combination of farmer behaviour, policy influence, and economic realities [69] .

Challenges faced with AI and hybrid models

As we know AI and machine learning is used by everyone and it has many powerful features, and AI models act as "black boxes". They can explain and predict information about biological modelling but not always correct data [70] . Integration of data-driven methods and process-based understanding in a clear and explained way remains a key difficulty.

Climate extremes and system disruptions

Due to extreme weather events such as drought, flood, heatwaves etc it is bit challenging for agricultural models to predict accurate data otherwise in normal weather conditions they provide good information about crop performance. Sudden change in weather or attack of pests and disease also causes losses up to large-scale which is sometimes not captured or observed by models. For betterment of future models' resilience and system recovery should be prioritized first, ensuring farms long-term adaptability instead of focusing on short-term yield changes [71] .

Balancing productivity and sustainability

As population is increasing day by day worldwide it is necessary to protect the resources such as soil, water, biodiversity, and livelihoods while producing food [39] . Enhancement of crop yield is not enough for creating sustainability in agriculture resources used in farming also needs proper attention [72] . Many models mainly focus on crop yield and disregard these topics. Sustainability still focused on developing models with multiple objectives which can further balances the productivity by keeping in mind the idea of environment protection and social welfare of people.

Conclusion

The agricultural and biological modelling has come a long way and evolved from simpler observations into sophisticated data systems to get accurate data of farming systems and its working. New agricultural models are designed with multiple objectives not only about crop performance and soil, but also includes the environmental processes and social well-being of humans. The integration of process-based models with artificial intelligence, remote sensing, and other high technology data which helps in providing accurate predictions and make it easy for farmers and policy makers prepares them to face conditions such as weather change and economic change. Development of models is not about better technology but still there are many challenges and issues in way of modelling. Sometimes there is limitations to real-farm data which is difficult for models to link across different scales. Climate extreme events such as droughts, floods, and heatwaves along with pest and disease attack which is bit challenging for models sometimes because sudden change in weather causes less accurate data prediction. Another important thing is yield is not only important thing but also the resources, their protection and sustainability is also very necessary to a great aspect. So, the future models are designed with the idea to conserve biodiversity, soil, water, and also farmers livelihoods. Models and high technology are the future which should be created by keeping everything in mind to create sustainability and save the environment. Agricultural and biological modelling should give more attention to people and living environment not only to technology. Models which are human-centred considers need of farmer, local knowledge, and real-world limitations, which makes it more reliable and easier to utilize. Transparency and ethical practices allow users to understand better about models output with reducing bias or misuse. These futuristic models can help providing better decisions to support resilient farming system with the collaboration among scientists, policymakers, and farmers, also keeping the idea of prioritizing sustainability, protecting the environment, resources and future livelihood.

References

1. Hatfield JL, Gitelson AA, Schepers JS, Walthall CL. Application of spectral remote sensing for agronomic decisions. Agron J. 2008;100(3 SUPPL). doi: 10.2134/agronj2006.0370c.
2. Ewert F, et al. Crop modelling for integrated assessment of risk to food production from climate change. Environ Model Softw. 2015;72:287-303. doi: 10.1016/j.envsoft.2014.12.003.
3. Challinor AJ, Watson J, Lobell DB, Howden SM, Smith DR, Chhetri N. A meta-analysis of crop yield under climate change and adaptation. Nat Clim Chang. 2014;4(4):287-291. doi: 10.1038/nclimate2153.
4. Schmidhuber J, Tubiello FN. Global food security under climate change. Proc Natl Acad Sci U S A. 2007;104(50):19703-19708. doi: 10.1073/pnas.0701976104.
5. Liakos K, Busato P, Moshou D, Pearson S. Machine learning in agriculture: A review. Sensors. 2018;18(8).
6. Kamilaris A, Prenafeta-Boldú FX. Deep learning in agriculture: A survey. Comput Electron Agric. 2018;147:70-90.
7. Keating BA, et al. An overview of APSIM, a model designed for farming systems simulation. Eur J Agron. 2003;18(3-4):267-288. doi: 10.1016/S1161-0301(02)00108-9.
8. Vapnik VN. The Nature of Statistical Learning Theory. New York: Springer; 2000. doi: 10.1007/978-1-47573264-1.
9. Cao W, Jeffert WW, Enli W. Crop modelling and decision support. 2004;(1):1-14.
10. Lobell DB, Burke MB. On the use of statistical models to predict crop yield responses to climate change. Agric For Meteorol. 2010;150(11):1443-1452. doi: 10.1016/j.agrformet.2010.07.008.
11. de Wit CT, de Wit CT. Photosynthesis of leaf canopies. Agric Res Reports. 1965;(663):1-54. Available from: http://apps.webofknowledge.com.
12. Monteith JL. Climate and the efficiency of crop production in Britain. Philos Trans R Soc London B Biol Sci. 1977;281(980):277-294. doi: 10.1098/rstb.1977.0140.
13. Dar NA, et al. Integrated farming systems for sustainable agriculture. In: Eco-friendly Agrobiological Tech Enhancing Crop Product. 2018. p. 111127. doi: 10.1007/978-981-10-6934-5_6.
14. Doraiswamy PC, Hatfield JL, Jackson TJ, Akhmedov B, Prueger J, Stern A. Crop condition and yield simulations using Landsat and MODIS. Remote Sens Environ. 2004;92:548-559.
15. Wang H, et al. Physics-guided neural networks for modeling dynamical systems. 2020.
16. Asseng S, et al. Uncertainty in simulating wheat yields under climate change. Nat Clim Chang. 2013;3(9):827832. doi: 10.1038/nclimate1916.
17. Burr T. Pattern Recognition and Machine Learning. J Am Stat Assoc. 2008;103(482):886-887. doi: 10.1198/jasa.2008.s236.
18. Lu ZQJ. The Elements of Statistical Learning: Data Mining, Inference, and Prediction. J R Stat Soc Ser A Stat Soc. 2010;173(3):693-694. doi: 10.1111/j.1467985x.2010.00646_6.x.
19. Reichstein M, et al. Deep learning and process understanding for data-driven Earth system science. Nature. 2019;566(7743):195-204. doi: 10.1038/s41586019-0912-1.
20. Crane-Droesch A. Machine learning methods for crop yield prediction and climate change impact assessment in agriculture. Environ Res Lett. 2018;13(11). doi: 10.1088/1748-9326/aae159.
21. Lobell DB, Hammer GL, McLean G, Messina C, Roberts MJ, Schlenker W. The critical role of extreme heat for maize production in the United States. Nat Clim Chang. 2013;3(5):497-501. doi: 10.1038/nclimate1832.
22. Holzworth DP, et al. APSIM-Evolution towards a new generation of agricultural systems simulation. Environ Model Softw. 2014;62:327-350. doi: 10.1016/j.envsoft.2014.07.009.
23. Assessment E. Assessing agricultural risks of climate change in the 21. 2013. p. 1-31.
24. Bouis HE, Saltzman A, Birol E. Improving nutrition through biofortification. In: Agriculture Improvement for Nutrition. 2019. p. 47-57. doi: 10.1079/9781786399311.0047.
25. Li Y, Li M, Li C, Liu Z. Forest aboveground biomass estimation using Landsat 8 and Sentinel-1A data with machine learning algorithms. Sci Rep. 2020;10(1). doi: 10.1038/s41598-020-67024-3.
26. Lundberg SM, Lee SI. A unified approach to interpreting model predictions. In: Advances in Neural Information Processing Systems. 2020. p. 4765-4774.
27. Ridley M. Explainable Artificial Intelligence. Res Libr Issues. 2019;(299):28-46. doi: 10.29242/rli.299.3.
28. Sudha SP, Loret JBS. A review on machine learningbased precision agriculture techniques for crop farming monitoring with IOT. Discov Environ. 2026;4(1). doi: 10.1007/s44274-025-00305-8.
29. Robert PC. Precision agriculture: A challenge for crop nutrition management. Plant Soil. 2002;247(1):143-149. doi: 10.1023/A:1021171514148.
30. Jones JW, et al. Toward a new generation of agricultural system data, models, and knowledge products: State of agricultural systems science. Agric Syst. 2017;155:269-288. doi: 10.1016/j.agsy.2016.09.021.
31. Verschuuren J. Agriculture, forestry and other land use (AFOLU). In: Research Handbook on Climate Change Mitigation Law. 2022. p. 433-456. doi: 10.4337/9781839101595.00025.
32. Rosenzweig C. Potential impacts of climate change on agriculture. In: A Spectrum of Achievements in Agronomy. 2015. p. 73-88. doi: 10.2134/asaspecpub62.c9.
33. Thornton PK, Herrero M. Climate change adaptation in mixed crop-livestock systems. Glob Food Sec. 2014;3(2):77-84.
34. Roscher R, Bohn B, Duarte MF, Garcke J. Explainable Machine Learning for Scientific Insights and Discoveries. IEEE Access. 2020;8:42200-42216. doi: 10.1109/ACCESS.2020.2976199.
35. Rockström J, et al. Planetary boundaries: Exploring the safe operating space for humanity. Ecol Soc. 2009;14(2). doi: 10.5751/ES-03180-140232.
36. Bongiovanni R, Lowenberg-Deboer J. Precision agriculture and sustainability. Precis Agric. 2004;5(4):359-387. doi: 10.1023/B:PRAG.0000040806.39604.aa.
37. Sehgal V. Remote Sensing for Crop Growth and. 1999;(i).
38. Kume A. Remote sensing of vegetation. Japanese J Ecol. 2014;64(3):201-204.
39. Gebbers R, Adamchuk VI. Precision Agriculture and Food Security. Science. 2010;327(5967):825-828.
40. Turner B. Land System Science: Land systems, sustainability and land system architecture. In: 2nd Global Land Project Open Science Meeting; 2014.
41. IPCC. Climate Change 2022-Impacts, Adaptation and Vulnerability. Cambridge University Press; 2023. doi: 10.1017/9781009325844.
42. Ostrom E. A general framework for analyzing sustainability of. Science. 2009;325(5939):419-422.
43. Van Ittersum MK, Cassman KG, Grassini P, Wolf J, Tittonell P, Hochman Z. Yield gap analysis with local to global relevance-A review. F Crop Res. 2013;143:417. doi: 10.1016/j.fcr.2012.09.009.
44. Wessels J, Wierzbicki AP. Model-Based Decision Support. In: Modeling and Decision Support Methodology. 2000. p. 9-28. doi: 10.1007/978-94-015-9552-0_2.
45. Vigo M. Precision Agriculture: Spatial and temporal variability of environmental quality. 1993;119:141-142.
46. Mulla DJ. Twenty five years of remote sensing in precision agriculture. Biosyst Eng. 2013;114(4):358371.
47. Atzberger C. Advances in Remote Sensing of Agriculture. Remote Sens. 2013;5(2):949-981.
48. Burrough PA, McDonnell RA. Principles of Geographical Information Systems. Princ Geogr Inf Syst. 1998;75(4):422.
49. McBratney A, Whelan B, Ancev T, Bouma J. Future directions of precision agriculture. Precis Agric. 2005;6(1):7-23. doi: 10.1007/s11119-005-0681-8.
50. Rouse JW, Hass RH, Schell JA, Deering DW. Monitoring vegetation systems in the great plains with ERTS. Third Earth Resour Technol Satell Symp. 1973;1:309-317.
51. Stafford JV. Implementing precision agriculture in the 21st century. J Agric Eng Res. 2000;76(3):267-275.
52. Corwin DL, Lesch SM. Characterizing soil spatial variability with apparent soil electrical conductivity: I. Survey protocols. Comput Electron Agric. 2005;46(13):103-133. doi: 10.1016/j.compag.2004.11.002.
53. Wheeler T, Von Braun J. Climate change impacts on global food security. Science. 2013;341(6145):508-513. doi: 10.1126/science.1239402.
54. FAO. The future of food and agriculture: trends and challenges. 2017;4(4).
55. Dixon J, Gulliver A, Gibbon D. Farming Systems and Poverty. FAO. 2001. p. 353.
56. Langdon S, Ritter ARM, Samy Y. Agriculture and Rural Development. In: African Economic Development. 2019. p. 268-302. doi: 10.4324/9781315690506-12.
57. El Bilali H, Bassole IHN, Dambo L, Berjan S. Climate change and food security. Agric For. 2020;66(3):197210. doi: 10.17707/AgricultForest.66.3.16.
58. Park A. Equity in Sustainability. USDN. 2014.
59. Bandura A. Social Foundations of Thought and Action. 1986. p. 617.
60. Gooch D. Communications of the ACM. XRDS Crossroads ACM Mag Students. 2011;18(2):6. doi: 10.1145/2043236.2043240.
61. Garnett T, Godfray HCJ. Sustainable Intensification in Agriculture. 2012. p. 23.