Hi Franck, I tinkered a little tonight - added some recent data from sampling locations close to the discharge point, changed the location of the discharge point (estimated by zooming in to max on https://www.monitororbs.jp/en/ - I will ask the Japanese for the actual coords - however as we have discovered even +/- couple of hundred metres could influence interpretation), and renamed radionuclides using a capital letter (H-3 instead of h3).

I think it would be useful to categorise sampling locations as follows: 1) Those close to the discharge point (DP) (<0.7 km with the new DP coords, maybe further if we exclude those in 2)) at which the influence of the discharges can be detected 2) Those close to the FDNPS port at which the influence of the continuing releases from the FDNPS site are evident and which more or less drown out any influence from the ALPS discharges. 3) All other sampling locations at which there is no H-3 above the regional background. This would have to be done manually I think. Also it is risky because if the magnitude of the discharges increases things will change (higher levels discharged, more dispersion required before levels drop to background, spatial extent of influence of discharges is bigger). But for now things are pretty constant so fixed categories should be fine. We could also try defining and selecting by polygons, i.e. one for each sampling location category?

Then to explain fully we should show the graph for results from each category of sampling locations plus a simple map showing from which sampling locations data has been selected (in my notebook there was a map showing all sampling locations with selected ones highlighted)

We could also port the activity by distance chart but additionally fill by sampling location category?