Abstract:
The acquisition of spatial data can be a problematic process, especially for geographically isolated areas or those where fieldwork is difficult. It is, therefore, important to explore non- field-based methods of producing geospatial data layers for such areas. Here the use of freely available resources and methods of producing geospatial layers are evaluated with the aim of producing basic basemap features such as contour lines, rivers, towns, and roads. These methods are statistically analysed and validated to ensure the accuracy of the features produced. Rodrigues island (Mauritius), is used as the study area, covering an area of 104 km2 with the highest peak (Mont Limon) reaching 396 m a.s.l. The island offers a dynamically varied terrain ranging from steep slopes to relatively flat coastal regions, allowing the methodology to be tested over all terrain types. Elevation points were produced using freely available resources, such as Training Center XML (TCX) Converter, and Terrain Zonum Solution. These were interpolated using GIS to create DEMs using two interpolation methods (Inverse Distance Weighted (IDW); Ordinary Kriging). IDW was chosen as a simple interpolation method, Ordinary Kriging as a more statistically robust method. The output DEMs were used as the basis for subsequent data extraction and creation. Hydrological modelling was used to model drainage lines; towns, roads, and dams were manually digitised using the freely available software Google Earth™ as the source. With statistical validation IDW proved to predict elevation values that correspond/correlate more with the elevation values of the control DEM, than those generated from the Ordinary Kriging. However, both methods returned outputs that closely resembled the control DEM and were deemed to be acceptable for data creation. Once all required geospatial layers were produced, they were compiled into a complete basemap and compared to the geospatial data collected by the Surveyor General of Mauritius. Although both maps were similar, multiple areas of differences were identified; these areas were ground truthed to determine and validate the findings. Ultimately it was determined that users can produce basemap features of sufficient accuracy for areas that either do not have geospatial data available or are difficult to access. As such, the framework proposed here may be followed to create basic geospatial layers for other inaccessible areas that exhibit similar geographic characteristics.
