Improving Snow Canopy Interception Modelling Using Airborne Lidar Data

Abstract:

Forest snow interception can account for large snow storage differences between open and forested areas. These differences also create large variations in sublimation, with estimates varying from 4% to 50% worldwide. This broad range highlights the need for accurate estimation of canopy interception.

Most current interception models utilize canopy cover and leaf area index (LAI) to partition interception efficiency (interception / precipitation in the open) which is typically modelled as an exponential decrease with increasing precipitation. However, these models can show limited utility quantifying snow interception dynamics under heterogeneous canopy. This study paired large scale field measurements of snow interception to aerial LiDAR data in efforts to better estimate interception within a Norwegian spruce dominated forest.

Seven field areas were equipped with precision geo-located sampling grids within various canopy density/gap fractionation regimes in three elevation bands surrounding Davos, Switzerland. Snow interception measurements were taken at each point for a series of storm events from 2012 to 2014 for a total of ~10,000 interception measurements and compared with further measurements at two open sites.

Existing and novel canopy metrics at the field areas were developed using a high resolution airborne LiDAR data set. These include estimates of LAI, canopy closure, under canopy incoming solar radiation, distance to tree measurements, aerial gap fraction measurements and various tree size parameters. These estimates were validated from 112 hemispherical images of the canopy structure within the field sites.

The snow measurements and LiDAR derived canopy metrics were then integrated to formulate an improved representation of canopy interception and further compared to interception estimates from the use of standard canopy metrics.

The standard mechanistic framework utilized in many models was also modified based upon interception efficiency distributions seen from this dataset. This model approach gives more accurate estimates of interception within heterogeneous canopy cover at various scales.