Evaluation of potential productivity in coniferous forests by integrating field data and aerial laser scanning in Hidalgo, México

Aim of study: To predict the productivity potential of a managed conifer forest by estimating the site index from Light Detection and Ranging (LiDAR) data. Study area: Intensive Carbon Monitoring Site Atopixco, Hidalgo, Mexico. Material and methods: A total of 329 observations from five remeasurements in permanent forest inventory sampling units were used to generate site index curves and metrics derived from a 2013 LiDAR scan. LiDAR elevation metrics were statistically related to field-observed dominant height (DH). Three models were fitted to predict DH as a function of LiDAR metrics, while nine height growth models were developed using the algebraic difference approach, at a base age of 40 years, using the ordinary least squares method and mixed effects models (MEM). Main results: The 99th height percentile was the LiDAR metric that showed the greatest correlation with the observed DH. Its integration into a linear model was best suited to estimate DH with Adjusted Determination Coefficient (R2adj) of 0.97 and Root Mean Square Error (RMSE) of 0.31 m. The Hossfeld IV anamorphic model adjusted as MEM and autocorrelation corrected model showed the best performance for predicting DH growth with R2adj of 0.87 and RMSE of 2.11 m. The integration of both models into a Geographic Information System (GIS) allowed the spatially explicit construction of an accurate mosaic of the DH and site index to classify stand productivity in the study area. Research highlights: Of the total area managed for timber purposes, 87% is classified as a heigh (≥31 m) and average (26 m) site index, while areas dedicated to conservation contain 13% of the area classified with low site index (≤21 m).