search

ornldaac / gedi_tutorials

GEDI L3 and L4 Tutorials
Other
115 stars 51 forks source link

Two questions about the tutorial #3

Closed sbanchero closed 3 years ago

sbanchero commented 3 years ago

Hi! I've two questions about the tutorial and the AGBD calculation.

1) The document [1] and [2] specify that par[1] is always the intercept term but in the tutorial 3_gedi_l4a_exploring_data.ipynb say: The variable par gives model coefficients - the first element of the list is always the slope of the linear model. Which is the correct one?

image

2) Also, I have another question regarding the use of the existing model in ANCILLARY / model_data. For example, in the DBT_NAm case if the vertors are: predictor_id: [1 2 0 0 0 0 0 0] rh_index: [50 98 0 0 0 0 0 0] par: [-120.77709198 5.50771856 6.80801821 0. 0. ]

How use the information to make the model equation? The vectors index starts from 0 or 1?

Follow both [1] or [2] is some confusing the order of the parameters because I don't understand how I should combine the rh values as only have two values 1 and 2 within the prediction_id for rh50 and rh98 respectively but tree par coefficients.

I understand that a possible way can be: AGBD = -120.77709198 + 5.50771856 rh50 + 6.80801821 rh98^2

Is this correct?

I hope my explanation has been understood.

Thank you very much for the tutorials. Best regards Santiago

[1] https://daac.ornl.gov/daacdata/gedi/GEDI_L4A_AGB_Density/comp/GEDI_L4A_Common_Queries.pdf [2] https://daac.ornl.gov/GEDI/guides/GEDI_L4A_AGB_Density.html

rupesh2 commented 3 years ago

HI @sbanchero , 1) par[1] should be the "intercept"; it is now updated in the tutorial description. 2) The equation for DBT_NAm should be AGBD = -120.77709197998047 + 5.507718563079834 x RH_50 + 6.808018207550049 x RH_98. predictor_id provides here mapping between rh_index and par. Also, note that both predictor and response variables, in this case, are transformed (sqrt).

sbanchero commented 3 years ago

Hi @rupesh2 , Thank you very much for your answer