Abstract
Sea level rise is a coastal hazard leading to erosion, flooding, and habitat destruction among other effects. Geoinformatics plays a vital role in providing tools to assess coastal vulnerability. Recent improvements in drone technology offer new opportunities for collecting data. Calibrating drone elevation data to local sea-levels is a current challenge. This research compares the use of drone-derived digital elevation models (DEMs) created from a DJI Phantom 4 Pro drone with airborne LiDAR data to analyze sea-level rise in Essex, Massachusetts. To evaluate the best method of calibrating drone elevation data to the local environment, three different Ground Control Point (GCP) methods were evaluated: surveyed GCPs, LiDAR-derived GCPs, and the NAVD88 sea-level (0-value) of the LiDAR data. The three drone-derived DEMs were compared with a LiDAR DEM through two methods: 1) comparing how well the DEMs measured the elevation of surveyed GCPs, and 2) how well the DEMs modeled five different scenarios of sea level rise compared with a LiDAR DEM. Results showed that two of the calibration methods performed well; the surveyed GCP derived drone DEM and the LiDAR NAVD88 sea-level (0-value) derived drone DEM.
Abstract
Snow cover has a major influence on the global energy balance through the reflection of shortwave solar radiation as well as influencing ecological processes and human activity. Numerous studies have found that snow cover extent (SCE) is decreasing in the Northern Hemisphere and this decline appears to be influencing temperatures and might be a major factor in the polar amplification. This research used satellite-derived Moderate Resolution Imaging Spectroradiometer (MODIS) snow cover data (MOD10C2), Land Surface Temperature (LST) data (MOD11C3), and Normalized Difference Vegetation Index (NDVI) data (MOD13C2) to detect changes in SCE and its potential relationship to changes in land surface temperature and vegetation growth from 2000 to 2017 over northeastern North America. There is a lack of detailed research concerning these variables for northeastern North America. The data were composited into seasonal and annual (snow-year: September–June) groupings. Two different change analyses were undertaken: 1) significant change using the Mann–Kendall statistical analysis and 2) univariate differencing using three different time periods (3 years, 5 years, 8 years). A regression and correlation analysis was undertaken between SCE and LST and NDVI to determine the relationship between changing SCE and changes in LST and NDVI. Based on the Mann–Kendall statistical change analysis (p-value = 0.05) for the 16-day data (32-day data), the area of declining SCE was more than 12 times the area of increasing SCE (more than 5 times for 32-day data) with declines occurring in all seasons, most notably in fall, June and the entire snow-year. Based on the univariate differencing analysis, SCE declined more than increased 96% of the time. Based on the regression/correlation analysis, SCE explains variability in LST (NDVI) for the snow-year: 43% (51%), spring: 31% (22%), June 34% (no significant relationship), fall: 40% (no significant relationship), and winter with no significant relationship (30%). It was determined that there is a weak to moderate inverse relationship between SCE and LST and a similar, but less prominent relationship between SCE and NDVI. A multiple regression/correlation with SCE and LST (independent) and NDVI (dependent), LST was a better predictor of NDVI than SCE. This relationship indicates that there is a potential positive feedback mechanism warming the region and increasing the region’s NDVI.