error messages:" Mult2 out of bounds in module gsland"

[coespicmbio]

I can’t use directional connectivity in the analysis. What possible problems can generate the following error messages: “showmessage: Mult2 out of bounds in module gsland = -7.45058e-09”? In output there is the information “Successfully loaded 1st stage of planning units connectivity hierarchy” but also “WARNING: BQP switched off as tree connectivity model is used...”.
If it helps below are the settings used:
[Settings]
removal rule = 1
warp factor = 1
edge removal = 1
use planning unit layer = 1
planning unit layer file = C:/Desktop/Zonation/unidade_planejamento/up_primha_23_09_19.tif
use condition layer = 1
condition file = C:/Desktop/Zonation/inputs/condition_layers_list.txt
use groups = 1
groups file = C:/Desktop/Zonation/inputs/groups_list.txt
use mask = 1
mask file = C:/Desktop/Zonation/unidade_planejamento/uc_primha_1.tif
use boundary quality penalty = 1
BQP profiles file = C:/Desktop/Zonation/inputs/nqpcurves1.txt
BQP mode = 2
BLP = 0
use tree connectivity = 1
tree connectivity file = C:/Desktop/Zonation/inputs/my_tree_file_2.txt
use interactions = 0

Thank you!

[PKullberg]

I haven't ever tried running a setup with both boundary quality penalty and tree connectivity on at the same time, but the source code and the error that you get seems to imply that it is not allowed. The "Mult2 out of bounds in module gsland" error might be simply a consequence of trying that. Have you tried a analysis set with only one of the connectivity features on?

As a side note, I took a quick look at the manual and it seems that this restriction is not mentioned there and should be corrected.

[coespicmbio]

I haven't ever tried running a setup with both boundary quality penalty and tree connectivity on at the same time, but the source code and the error that you get seems to imply that it is not allowed.
Have you tried a analysis set with only one of the connectivity features on?

Hi PKullberg thanks for your quick reply!

We have indeed tried a variety of iterations and changed the different settings with each run, but the message still persists. So one run was without boundary quality penalty and we even ran with edge removal turned off to see if it changed anything.

We suspect that the message might not be a "hard-error" per se but one where the algorithm just can not find a solution using the connectivity tree, because the process jumps to the next PLU.
However, given the size of our study area (whole Amazon basin) the machine starts to demand a lot of the CPU memory, so it might be a capacity limitation rather than problem with the logic.

What could we do to help you analyse this?