A low-cost alternative to LiDAR for site index models: applying repeated digital aerial photogrammetry data in the modelling of forest top height growth

Abstract Environmental and forest structural information derived from remote sensing data has been found suitable for modelling forest height growth and site index and therefore forest productivity assessment, with the advances in airborne laser scanning (ALS) playing a major role in this development. While there is growing interest in the use of ALS-derived point clouds, point clouds from high-resolution digital aerial photography (DAP) are also often used for mapping and estimating forest ecosystem properties due to their lower acquisition costs. In this study, we document the applicability of bi-temporal DAP data for developing top height (TH) growth models for Scots pine stands. Our results indicate that DAP data can function as an alternative to traditional TH measurements used in growth modelling when corrected based on a limited sample of field-measured reference TH values. As the correction cannot be constant for each DAP dataset due to the different parameters during data acquisition, we propose a straightforward method for the bias correction of DAP-derived TH estimates. By undertaking iterative random sampling, we were able to find the minimum number of reference measurements needed to calculate the TH correction in order to achieve the desired accuracy of the TH estimations based on DAP. Here, we used ALS data as the reference data; however, the ALS measurements can be replaced by any other reliable source of TH values. The presented method for determining TH can be used not only for site index and forest growth modelling but also in forest inventories.