National Scale land use Change Modeling – Issues and Applications
An example in modeling:
A framework of the conceptual model for land use change evaluation at the national level and its forecasting is shown in Fig. 1. The issues in modeling are discussed with respect to this framework. Table 1 Lists the data that are used in the model and the minimum required dataset resolution and classification schema required.
Fig. 1. A Simple Framework for Land Use/Cover Change Modeling
Table 1. Data required in National scale Land use Modeling
|Type of Data||Data Characteristics(min. requirements)||Remarks, if any(Source, limitations, etc)|
|1. Climatic data2. (rain and temperature)||Point data – have to convert to spatial (grid) data- denser the meteorological stations, the better the interpolation.||Number of WMO stations are limited|
|2. (a) Elevation data||Atleast 1km DEM data is needed||Available in GTOPO dataset.|
|2.(b) landuse data||1km resolution data or digitization of scales higher than or equal to 1:1 million||Classification schema should atleast have the major classes including some of the major crops|
|3. Population||Urban-rural breakup at local administrative units||In case of Thailand, data at province level.|
|4. GNP & Income Levels||Available only at the national level, used as an index of affuluence|
|5. Costs of Production and Price of Products (Agricultural)||Non-spatial, so data at the lowest administrative level best.||Depends on the data collection agencies.|
The crop-wise productivity estimation requires a set of physical and climatic data. The data on rainfall and temperature is of prime importance in estimating the rain-fed potential of a given area. The better the precipitation data, the easier it will be to demarcate the growing seasons of different regions. In our study, initially we sourced our data from the Global Daily Summary (1994), which gave us a set of 62 stations for Thailand. The reliability of the interpolated precipitation data was not very good. So, we collected data from all the rain-gauge stations in Thailand, a set of 517 stations, and then created the rainfall map. It is clear from Fig. 2. that atleast a decent spread of meteorological points are needed to derive the rainfall distribution.
Fig. 2. Rainfall distribution of Thailand for the month of August 1991 (Data is interpolated from – Left: 517 rain-gauge stations and Right: 62 stations)
Similarly, the elevation data is of importance in deriving a lot of variables of the model, such as potential evaportanspiration, transportation coasts, etc and hence needs to be accurate enough to a certain degree. We adopted the method of deriving the DEM data from digitized contour maps, as there was no digital DEM data available. The digital elevation model was derived from 100 m contour lines by applying spline function. The DEM that we arrived at had some anomalies, especially in the central and north-eastern plains of Thailand. The recently published one-km DEM data from the GLOBE project (1997), shown in Fig 3., is a much better dataset, which helped us overcome some of the errors that we had encountered earlier.
Fig. 3. Digital Elevation Model of Thailand, reclassed for this figure (Left: from GLOBE project; Right: interpolated from 100m contour data)
Also, of importance is the Land use/cover data. The resolution and classification of the land use/cover data is of great significance when used for validation of the model results. If one has a data which delimits only the major landuse categories like paddy, crop, forest, urban, etc., then it is difficult to validate the simulationof changes withing that landuse like changing from data handling and proper reclassification. Fig 4. shows the two different sets of landuse data that was obtained from the relevant sources.
Fig. 4. Land Use maps of Thailand, 1990 (Left: Map showing 193 classes; Right: Map showing only 13 classes
Among non-spatial data used in land use/cover modeling, socio-economic data like population, and GNP are the major ones. GNP is used as an index of affluence and gives information on the consumption patterns of a country. While considering such data, it is necessary to develop scenarios and provide the possible combinations, of effects under these various conditions. Population distribibution and its breakup (rural-urban, agricultural labourers, sex) is also of importance in combining the social component into the model. The level at which these figures are available, will make a lot of difference in the effects generated due to shifts in these.
Application of Land use Models
Most of the global physical and biological models that have been developed for the understanding of the various parts of the global earth system, are dependent on the land use and land cover information as these with their control on the albedo and water & nutreient cycling establish the boundary conditions to these models (Robinson, 1994). Changes in land use/cover bring fundamental changes in the earth system, and these changes are unable to be accounted for by the global models, as they lack the ability to generate any comprehensive land use projections for linked to any such land use predictive models. Such a lack of the framework for the land use/cover trends means that the cause for changes are understood at a level, that is too general to provide any guidance to policy formulations at the national level and research priorities.
In a discussion on the effects of climate change on agriculture and future impacts on global food system, Parrry et. Al. (1994) show that the use of specific national model are required, in combination with the regional scale models to capture the true dynamics of changes that may result in the supply of food. Also, for arriving at major initiatives at the national and regional levels to protect the environment and to proceed towards sustainability, it is necessary to have models at the national-scale to give concrete information on the possible causes. Also, models with a flexibility towards scale of the domain of analysis will go a long way in providing appropriate information for planning at the various levels.