Land Cover Classification for Jambi Province, Indonesia using two different remote sensing techniques: visual interpretation and supervised classification

1. Justification
By date, land-use/land-cover changes for Jambi Province for the periods of 1990-2000, and 2000-2011 have been studied and show drastic and rapid dynamics. (Melati, D., et al, 2014). The same study states that the decrease of secondary forests was mainly due to the conversion into agriculture systems (oil palm plantation, rubber plantation, and mixed dryland agriculture) and also due to the establishment of timber plantation.
Considering the rapid transformation from forest to rubber and oil palm plantations in Jambi province, an updated land use/land cover map should be produced to monitor the changes and build a basis for decision making.

2. Objectives
Perform a new land use/land cover map for the Jambi Province, Indonesia based on the most recent Landsat data with a spatial resolution of 30 m.
This proposal follows the further research questions:
- Are agriculture systems increasing in the last years in Jambi province?
- How many hectares or percentage of land-use change is registered by date?
- Which areas are more affected in the province by these changes?

3. Methods
- Preparation/Data revision
o Check for available Landsat/Sentinel-2 data since 2013
o Use every possible and available image for creating a mask of non-interesting/non changing areas. E.g.: cities, oil palm plantations, lakes, etc.
Possible sources include Google earth, open source data, and former classifications
o Prepare a field manual for training data collection (including selection process of location, variables and descriptions)
- Fieldwork
o Visit different locations according to the previous revision of data according to field manual
o Characterize the locations according to field manual
o Make corrections if needed
- Classification
o Pre-processing of Landsat data: general preprocessing steps for visual and shortwave infrared (vis-SWIR) data
o Classifying
o Validation