LAND USE/ LAND COVER CHANGE MODELLING: ISSUES AND CHALLENGES

Land use change modelling are the tools to support timely and effective monitoring of natural resources through spatio-temporal land use/ land cover (LULC) change detection which can help decision makers for optimum resources planning and utilisation for sustainable rural development. Number of models and approaches have been developed in recent past to analyse land use/land cover change considering different type of data/variables at different levels of complexity and resolution. Such different methods/ models have their own limitations, advantages and suitability in a particular condition. There is no agreement among research community about suitability and effectiveness of a particular method. Present study aims to present a comparative study of popularly used LULC change detection models and techniques. Different models and techniques are compared in terms of level of complexity, considered explanatory variables, spatial extent, temporal dynamism,predictability and level of spatial interaction. For each category, a thorough review of models and approaches is presented which helps in understanding the operational concepts and utility of models/ approaches.


Introduction
Rural development requires effective monitoring and inventory of natural resources like land, vegetation, forest, water, agriculture, etc., for their optimum planning and utilisation. Land use/ land cover (LULC) change is one of the innovative techniques for natural resources inventory, monitoring and management.
Development leads to landscape disturbances, and alteration in LULC and natural resources.
Global earth system is bound to remain balanced without which it may produce some irreversible impacts on climate, ecosystem, agriculture system, biodiversity and environment (Jetz et al., 2007;Akin et al., 2014). Therefore, for sustainable development, feild data collection, spatial and temporal monitoring, LULC change detection, methods and models are effective tools. Over the last few decades, a range of LULC change models and approaches have been developed to accurately assess land transformation and functioning of earth system (Agarwal et al., 2002;Jat et al., 2008;Saxena et al., 2016). Land use change modelling, especially if performed in an integrated, multi-scale, dynamism and spatially explicit manner, can be proved as an essential component for land use change forecasting (Aspinall, 2004). An accurate forecasting may enhance the understanding of LULC change processes, driving factors and also can project alternative land use scenarios which may lead to the goal of sustainability. Models have the capability of representing a complex LULC system in an efficient way by incorporating multi-variates.
It also offers the geographical sensitivity and spatial suitability into the model which are utmost important for analysing LULC change for a region (Batty, 2005;Jat et al., 2017) (Agarwal et al., 2002;Baker, 1989). Hence, if becaomes still challenge ot identify best suitable model based on accuracy of simulation and spatial and temporal resolutions (Parker et al., 2003;Jat et al., 2008 (Xiubin, 1996).
However, for incorporating intensification into stochastic models, some efforts have been put which are helpful to be effective on LULC change intensification.

Optimisation Models
Optimisation is formed by combining small chunks analysis using differential equations. The prior understanding among variables is developed before simulation (Waddell, 2002). This modelling approach is more powerful in establishing the relationship among variables and simulating the dynamic behaviour of the system. Although, the relationship cannot be built into the system in a straightforward way, analysing LULC change with this method is complex as numerous interactions among variables take place.

Cellular Automata-based Models
Cellular automata (CA) is a cell-based framework which includes cell states, transition rules, infinite discrete cells and neighbourhood.
CA-based models are very effective for addressing LULC change. Urban growth is a dynamic and complex phenomenon and CA includes the capability to model a complex phenomenon in an easy and effective manner (Clarke and Gaydos, 1998).

Integrated Models
Above discussion gives information about various types of LULC modeling approach. The recent advancement includes the integration of more than one approaches for providing added attributes to the LULC change modelling. Such modelling approaches are used where spatially explicit interaction and dynamic behaviour including long-term prediction, is required.
However, it is very complex to hybrid two different modelling approaches for a large landscape as the number of interactions greatly increase (Lambin et al., 2000).  However, Cellular Automata-based approaches were found to be promising in modelling of LULC changes, which is very much evident from the number of its applications reported in the literature in recent past (Agarwal et. al., 2000).