Spatial Scale and Land Use Fragmentation in Monitoring Water Processes in the Colombian Andes
Land use spatial heterogeneity limits the applications of point and coarse spatial resolution remote sensing data in the regional representation of environmental processes. In the Colombian central cordillera little research has been conducted at the watershed level to assess the spatial heterogeneity of the landscape and the effects of land use in the spatio-temporal variations of water related processes. This paper evaluates the spatial heterogeneity of a 940 ha site in the Colombian Andes, discussing the feasibility of using point and coarse spatial resolution remote sensing data in producing regional representations of hydrological processes in this landscape. The spatial heterogeneity of vegetation cover and agriculture systems of an agriculture watershed in the mid-range of the central cordillera (1000–2000 m.a.s.l) is analyzed. Descriptive statistics and Analysis of variance (ANOVA) were used to compare landscape fragmentation within and among three spatial units, 0.5, 1, and 3 km2, equivalents to pixel sizes commonly present in environmental satellite data. The GIS-Remote Sensing analysis showed a heterogeneous landscape rich in agriculture systems with small parcels of monocrops interwoven with six other agriculture systems including different degrees of intercropping and crop association. The results found an average fragment size of 9 ha, equivalent to a 300 m pixel with up to 10% fragments smaller than 0.5 ha (70 m pixel). No one of the seven land uses identified cover more than 50% of any of the three spatial areas under investigation. The results suggest that coarse spatial resolution remote sensing data or single site point data may be unsuitable for water studies for this landscape because different combinations of vegetation cover, agriculture systems, and topography are expected to produce different outcomes of hydrological process. In this case, single site point data or coarse spatial resolution remote sensing data are expected to oversimplify the local environment and produce limited assessments of regional water processes in this landscape.