The application of geospatial interpolation methods in the reconstruction of Quaternary landform records

Abstract Erosional landform features and their associated sedimentary assemblages (river terraces) often provide important records of long-term landscape evolution. However, the methods available for spatial representations of such records are typically limited to the generation of two-dimensional transects (valley long profiles and cross sections). Such transects limit the full quantification of system responses in a three-dimensional landscape (e.g., the identification of spatial changes in net sediment flux within a hydrological basin). The purpose of this paper is to explore the use of geospatial interpolation methods in the reconstruction of Quaternary landform records. This approach enables more precise quantifications of terrace landform records at a range of spatial scales (from a single river reach to geological basin scales). Here we use a case study from the Tabernas basin in SE Spain to test the applicability of multiple methods of geospatial interpolation in the reconstruction of Quaternary landforms (river terrace and alluvial fan remnants). We take steps in (1) refining the terrace data sets and the methods of technique application in order to reduce modelling errors, and (2) in highlighting the requirements for an assessment of interpolation method suitability when modelling highly fragmented landform records. The results from our study show that the performance of interpolation methods varies considerably and is dependent upon the data modelled. Method performance is primarily controlled by the inherent geomorphological characteristics (surface morphology and elevation) of the data; however, the attributes of data structure are significant. We further identify the importance of predefined model parameters (e.g., search radius) upon technique performance, increasing the appreciation of these commonly neglected variables in such studies. Ultimately, the overall applicability of the interpolation process is evidenced by the close correlation of surface volume data generated by all interpolation methods. These data would suggest that the interpolation technique can be applied in many forms as a useful tool in the reconstruction of Quaternary landform features.