Instead of doing supervised classification to produce a land use map from satellite images, can I just use an existing land use map (in shapefile format) to predict land use change(Just convert the shapefile to raster then undergo land use prediction)? Furthermore, does the land use type need to be general? For example, the land use consists of forest, agricultural, built-up and water bodies only. What if I want to be more specific like breaking down agriculture to several types like husbandry, palm oil and aquaculture?
Land use types depend on your purpose. You can collect field data for the different land use types and do the accuracy assessment. For accuracy assessment, the classified map compares with ground truth data.
Hi Firdaus,
Yes, you can do that. If you are looking for a visual qualitative analysis of the LU changes then you can compare the LU maps from different timelines and calculate the area wise changes. But if you want to look at the statistical and more technical changes, a comparison of different values from satellite imageries by producing FCC maps is always recommeded.
In case you want to validate the existing LU changes, then you will have to go for a site survey which is a hectic job (until you have the manpower and time required). In that case, also, satellite imagery can give you the latest update. Hope that helps.
Thanks!
You have really nice models doing that, but you need trends in population growth, employement, demand for each type of lands, scenario...
See for example:
- http://www.smartpop.be/ > see "Land Use and population model to 2050" bookmark
- Article Long-term simulation of urban development in support of spat...
It depends who made the land use map, and the various aspects of data quality, and more important if you intend on using a time series, if all the datasets have a consistent methodology. For instance you might have 5 datasets, but halfway through they split some categories and added others. Scale is important also. And how many categories in the classification - easy if it's binary or trinary, like impervious surface, agricultural, and natural, but we have something like 32 in our county, USGS has https://www.fs.usda.gov/detail/r5/landmanagement/resourcemanagement/?cid=fsbdev3_048169 for the nation - because your looking for transitions between pairs of uses.
Michael John Patrick so it can be done if the land use map is made in a consistent way? The land use map that I have is obtained from a department that is specialized in GIS and land use. As for categories in the classification, the land use map consists of many attributes but it's broken down into 3 categories: 1) General land use like forest, agriculture, water bodies and industry, 2 and 3 are more specific like types of industry etc.
Can I still use the classification and attributes that are more specific?
Of course you could. The most important consideration is that the polygons of traditional land use maps, if made without aerial photography and/or satellite RS represent extrapolation of (possibly scarce) point data. When using RS, the cartographic data are continuous and contiguous. In other words, you have a value available for each pixel.
Yes, you can use an existing land use map to predict land use change. However, the accuracy of the prediction may depend on the quality and resolution of the existing map and the accuracy of the predictions may decrease as the time lag between the existing map and the present time increases.
Regarding the level of specificity in the land use categories, it is possible to have more detailed land use categories, such as breaking down agriculture into different types. However, the level of detail should be based on the purpose of your analysis and the availability of data. If you have the necessary data and resources, a more detailed classification can provide a more accurate and meaningful analysis. However, if you lack the data or resources, a more general classification may still provide useful insights. It is important to balance the level of detail with the practicality of the analysis.
- Badges
- Science topic
- Similar topics
- Phytochemistry
- Extracts