Page 1 of 1
GEDI Canopy Height Accuracy/Interpretation over Savannah Woodlands and Open Grasslands
Posted: Thu Dec 03, 2020 11:31 am America/New_York
User Question: So I have a question about the GEDI data. I recently came across the GEDI instrument and as a RnD exercise have downloaded both Levels 2A and 2B over an area that I am familiar with, open savannah woodland, but the data is a bit disappointing in terms of accuracy/consistency. Now I understand that the dataset has been globally calibrated; is there a workaround to calibrate more accurately to local conditions? Are you able to give me some guidance on this or point me towards someone who can?
Basically the data tells me that open grassland has a canopy height in the region of 3.5m. I need to try and better understand the processing and how to best interpret the data to give me results that I can rely on. Is this 3.5m constant across the board, does it only apply to grassland, how reliable are the canopy woodland heights?
If you could point me in a direction it would be much appreciated.
Re: GEDI Canopy Height Accuracy/Interpretation over Savannah Woodlands and Open Grasslands
Posted: Thu Dec 03, 2020 11:32 am America/New_York
Science Team Response: In this case it’s important that we communicate the L2A relative height metrics are not direct measurement of canopy of tree height, but “lidar perceived” height metrics. Ground and canopy signals in short statured vegetation are often mixed due to the large pulse width of GEDI, which blurs these signals together. Bare ground may have RH100 values around 3m as a result of this (other factors such as slope will also influence).
The are a couple of new science team projects that are evaluating GEDI L2 products specifically for savannas and seeking to improve: (i) filtering of bare earth waveforms from those with vertical canopy structure; and (ii) improved signal processing for separating canopy and ground returns.