High-resolution tree-canopy change dataset for New York City, encompassing the period 2010-2017. This represents a top-down, 2D mapping of tree canopy change. Polygons are assigned to one of three classes: (1) No Change, (2) Gain, and (3) Loss. No change indicates that this portion of the canopy has undergone no modifications during the time period. Gain indicates that new tree canopy has appeared during the time period. Loss indicates that this portion of the tree canopy was removed during the time period.

As part of the 2017 NYC LiDAR/land cover project, a comprehensive, 8-class land cover map was developed for the city: (1) Tree Canopy, (2) Grass\Shrubs, (3) Bare Soil, (4) Water, (5) Buildings, (6) Roads, (7) Other Impervious, and (8) Railroads. This dataset was derived from a combination of LiDAR, imagery, and ancillary vector datasets. The primary source data for the Tree Canopy class was 2017 LiDAR (QL1 specifications). Object-based image analysis techniques (OBIA) were used to extract tree canopy (and the other land-cover classes). OBIA systems work by grouping pixels into meaningful objects based on their spectral and spatial properties. Within the OBIA environment, a rule-based expert system was designed to effectively mimic the process of manual image analysis by incorporating the elements of image interpretation (color/tone, texture, pattern, location, size, and shape) into the classification process. A series of morphological procedures were used to ensure that the end product was both accurate and cartographically pleasing. The dataset was also manually reviewed and corrected to eliminate non-systematic errors that could not be addressed by additional model refinement. The overall accuracy of the 2017 land-cover map was 98%; user's accuracy of the Tree Canopy class was 99%. To assess how the city's tree canopy had changed during the previous 7 years, the Tree Canopy class from the 2017 land-cover map was compared directly to tree canopy mapped as part of a UTC assessment in 2010. This comparison was possible because similar OBIA techniques and source LiDAR were used to map tree canopy in 2010.

Nonetheless, some methodological differences were inevitable because better data and OBIA modeling protocols were available in 2017. Accordingly, the change-detection analysis was carried out using an OBIA system that evaluated the validity of individual gains and losses, eliminating very small changes and also identifying erroneous change attributable to errors in the 2010 tree canopy layer. Additional errors were fixed during a comprehensive manual review of a preliminary change map.

High-resolution tree-canopy change dataset for New York City, encompassing the period 2010-2017. This represents a top-down, 2D mapping of tree canopy change. Polygons are assigned to one of three classes: (1) No Change, (2) Gain, and (3) Loss. No change indicates that this portion of the canopy has undergone no modifications during the time period. Gain indicates that new tree canopy has appeared during the time period. Loss indicates that this portion of the tree canopy was removed during the time period.

As part of the 2017 NYC LiDAR/land cover project, a comprehensive, 8-class land cover map was developed for the city: (1) Tree Canopy, (2) Grass\Shrubs, (3) Bare Soil, (4) Water, (5) Buildings, (6) Roads, (7) Other Impervious, and (8) Railroads. This dataset was derived from a combination of LiDAR, imagery, and ancillary vector datasets. The primary source data for the Tree Canopy class was 2017 LiDAR (QL1 specifications). Object-based image analysis techniques (OBIA) were used to extract tree canopy (and the other land-cover classes). OBIA systems work by grouping pixels into meaningful objects based on their spectral and spatial properties. Within the OBIA environment, a rule-based expert system was designed to effectively mimic the process of manual image analysis by incorporating the elements of image interpretation (color/tone, texture, pattern, location, size, and shape) into the classification process. A series of morphological procedures were used to ensure that the end product was both accurate and cartographically pleasing. The dataset was also manually reviewed and corrected to eliminate non-systematic errors that could not be addressed by additional model refinement. The overall accuracy of the 2017 land-cover map was 98%; user's accuracy of the Tree Canopy class was 99%. To assess how the city's tree canopy had changed during the previous 7 years, the Tree Canopy class from the 2017 land-cover map was compared directly to tree canopy mapped as part of a UTC assessment in 2010. This comparison was possible because similar OBIA techniques and source LiDAR were used to map tree canopy in 2010.

Nonetheless, some methodological differences were inevitable because better data and OBIA modeling protocols were available in 2017. Accordingly, the change-detection analysis was carried out using an OBIA system that evaluated the validity of individual gains and losses, eliminating very small changes and also identifying erroneous change attributable to errors in the 2010 tree canopy layer. Additional errors were fixed during a comprehensive manual review of a preliminary change map.