This paper provides a continuation of the results and efforts to continuously maintain and use a fleet of 120 detailed Building Energy Models (BEM) ofthe Sandia National Laboratories New Mexico and California sites. The fleet has continued to be used in new applications beyond its 1st round of sitewideenergy retrofit and climate assessments in 2014-2017. These include resilient energy systems assessments in 2018-2019 and institutional peak electricload characterization in 2020. The most recent work is a 2nd round site-wide energy retrofit assessment that is being planned. This paper shows the 10step procedure planned for this 2nd assessment and contrasts it to the 1st round of institutional energy retrofit analyses. The procedure involves calculatingdifference metrics between the various steps in the procedure that highlight the accuracy of the energy retrofit decisions being made. Here, energy retrofitdecisions involve deciding what specific building and energy retrofit is the next best choice based on metrics such as total energy saved, carbon offset, or energycost savings minus the energy retrofit implementation cost. The first difference metric delta 11 assesses the robustness of energy retrofit decisions with respect tohistorical, climate change, and extreme event weather futures, the second delta 15 assesses the robustness of energy retrofit decisions by comparing results beforeand after BEM calibration. This provides information that helps to show if the retrofit is very sensitive to other BEM input parameters that are alsouncertain. The third metric delta 19 involves empirical validation of energy savings or other metrics used based on actual metered results. A demonstration ofcalculating delta 11 shows how important climate and future weather is to energy retrofit decisions for a 96 BEM study with 2 energy retrofits involving roofinsulation and external wall insulation. Weather files for 2017-2020, Typical Meteorologic Year 3 (TMY3), and 3 extreme event scenarios were includedfor different weather futures. The results show that variations in energy savings are significant and the optimal decision set with baseline year 2020 is onlystable to 30 decisions of the 192 potential energy retrofit decisions. This shows that using a single weather future is likely to lead to sub-optimal choices.