Quantifying change in urban tree cover in the city of Lubbock, Texas, using LiDAR and NAIP imagery fusion

Abstract

Urban tree cover provides important ecosystem services, such as the maintenance of biodiversity, the conservation of water, and human health. An accurate estimate of the canopy cover of urban trees is essential to quantify spatial variations, monitor changes, and provide spatial prioritization for the expansion of urban tree cover. Mapping canopy cover in semi-arid urban landscapes using general-purpose light detection and ranging (LiDAR) presents significant challenges due to the effect of seasonality and aridity on tree crown structures and leaf phenology. We present a locally adapted method based on data fusion and variable window filtering. We fused airborne LiDAR (2011 and 2019) with near-concurrent National Agriculture Imagery Analysis (NAIP: four bands) orthophotos and developed a segmentation model [(Canopy height model: CHM >2m) & (Normalized difference vegetation index: NDVI >0.3)] to account for changes in urban tree canopy cover in Lubbock City (320.75 km²), Texas, United States. Our results indicate a 35.4 % increase in city canopy cover between 2011 and 2019, ∼3.86 % yr⁻¹, with the Northwest and Northeast quadrants showing the largest gains (45.2 % and 42.7 %, respectively). The validation of the resultant tree canopy map was performed using Wayback images from the Earth System Research Institute (ESRI). Canopy delineation improved with increased LiDAR pulse density and an increase in resolution of NAIP imagery in 2019. Our results demonstrate that LiDAR-NAIP fusion can capture canopy change in heterogeneous, semi-arid landscapes and highlight the need for spatially and temporally aligned LiDAR and NAIP acquisitions to support reliable tree inventories and ecosystem service assessments.