Cost-effective land cover mapping methods

Details

The Buyer wishes to investigate options for cost-effective land cover mapping over New Zealand to meet increasing needs for more timely snapshots of how our land cover is changing. This will support more frequent updating of environmental reporting indicators; more timely monitoring of changes in carbon stocks in the landscape; and an improved understanding of erosion susceptibility and the risk of increased sediment loads in waterways and estuaries. Currently land cover is mapped into the Land Cover Database (LCDB) approximately every five years. Thirty-three cover classes are delineated with a minimum mapping unit of one hectare. Updates to the LCDB map series are time consuming and costly.

The Buyer would like to investigate opportunities to leverage new technology and data to find more cost-effective and timely ways the map land cover over New Zealand, initially focusing on automated mapping of a generalised set of land cover classes, possibly to a finer scale (smaller minimum mapping unit) than LCDB. Ultimately such methods may be leveraged in the production of LCDB updates, if they can be integrated in such a way that the integrity of the LCDB time series is maintained.

The Buyer would like to test the utility of currently available global land cover data sets over New Zealand to identify the strengths and weaknesses of these products, both as a single snapshot and in terms of their ability to detect real land cover change through time. The Buyer also requires a cost-effect method to generate simplified (10-12 class) land cover maps over New Zealand, which provide accurate and consistent land cover classifications such that they would be suitable for monitoring change over time. This solution needs to be based on open, regularly updated imagery, and, as far as possible, on open-source code and systems.