Automatic harmonization of heterogeneous agronomic and environmental spatial data

摘要: The analysis and mapping of agronomic environmental spatial data require observations to be comparable. Heterogeneous datasets are those for which the different cannot directly compared because they have not been collected under same set acquisition conditions, instance within time period (if variable interest varies across time), with consistent sensors or similar management practices impact measured value) among others. When heterogeneous conditions take place, there is a need harmonization procedures make possible comparison such observations. This details compares four automated methodologies that could used harmonize agricultural so can analysed mapped conjointly. theory derivation each approach, including novel, local approach given. These methods aim minimize occurrence discrepancies (discontinuities) in data. approaches were evaluated sensitivity on simulated known characteristics. Results showed none consistently delivered better accuracy. accuracy preferred choice was shown influenced by (i) within-field structures datasets, (ii) differences between (iii) resolution real grain yield discussion help users select an appropriate methodology proposed. Despite significant improvements dataset harmonization, discontinuities entirely removed some uncertainty remained.