IDFxh1/102 tree decline and conversion to grasslands.

It shows both Py and Fd are declining over time until site is unsuitable by 2100. I question this - both Py and Fd do occur on rock outcrops in drier and hotter BGC units. Future BGC units include other IDFs and BG units. (BGmw_WA). The edatopic grids for these American grassland units lack forested rock outcrops on their edatopic grids. They only show grasslands on dry sites which is therefore the only option for this site series to shift to in those BGC units which is incorrect. Edaphic features - shalllow soils, exposed bedrock etc. control what vegetation will occur on these sites. The future BGC units may be correct but they lack the equivalent site series that would be expected on this site type.

Start by reviewing the IDFxh1/102 tree decline and the edatopic grids for the American grassland units, then compare the future BGC units with the rock-outcrop site conditions described here. Done means determining whether equivalent forested rock-outcrop site series are missing and documenting the appropriate correction.

Written by the indexing model from the issue text.