Compare TRIM to eDTM

[gcperk]

Compare the ensemble DTM to current data sources:

1. TRIM at a province levels

• Rough comparison at province wide (simple
• Hot spots

2. TRIM at known problematic areas

• lakes - rmse value (?).
• cross sections/ plot the cross sections. (lakes etc).
• cutblocks/ fires
• highly vegetated areas
• highly variable mountainous areas

3. For Lakes :

• lakes <- read in lakes from bcdata
  calc stats for elevations in lakes in both DEMs.
• Which DEM has a range closer to 0m?
• graph the elevation range of each lake to ID outliers and compare the spread between the two DEMs visually.
• ID lakes with elevations differences outside an acceptable threshold in both DEMs

[gcperk]

Review of data for Deception AOI. A comparison of the eDTM, trim and Lidar dem - showing two transect within Deception (location on which isat points fall.

Trim is consistently closer to TRIM as compared to eDTM

[gcperk]

Lidar subtract eDTM (left) and TRIM (right) for Deception AOI. Range of colours from red (-ve to green/blue (+ve). transparent is where delta is within 5m (-ve or +ve) ]

Overall thoughts

• TRIM estimates showed larger proportion of data with -5m to +5m than eDTM.
• eDTM showed bias to under estimate of elevation.
• eDTM sensitive to cutblocks
• both TRIM and EDTM showed higher error in areas with high hillshade.
• eDTM outperformed TRIM in some areas.