Recently on Quora someone asked: Different types of energy production create different types of environmental harm – how can we compare these varying types and degrees of environmental harm in an apples-to-apples way? My answer is copied here.
There are several mechanisms for cross-generation environmental impact assessment.
Meta-analyses of CO2e
The most important is meta-analysis of full-lifecycle CO2e-emissions, as global warming is the most critical environmental threat facing humanity. A meta-analysis gathers all Lifecycle Cost Analyses (LCA) for all generation forms for which LCAs are mandated for manufacturers and is conducted according to ISO standards. LCAs must be published  and subject to validation. A complete meta-analysis site is publicly available.  Here is an example chart created from this site showing CO2e per KWh for wind, nuclear, natural gas and coal generation. (Note that this meta-analysis pre-dates discovery of significant methane leaks at some shale gas fracking sites, which could negate the CO2e advantage over coal; the jury is out on that one at this point. )
Life Cycle Assessments by GE Research
General Electric manufacturers every form of generation technology. As such, it is required to create LCAs (once again, ISO standard, published and subject to audit) for every form of generation technology and that data is available internally, so when they publish comparisons, they are worth looking at. In this case, energy payback ratios are a surrogate for total pollution, but not specific pollution. In general, manufacturing, distribution and construction are typically heavily dependent on fossil fuels in most jurisdictions, so it’s a reasonably proxy for fossil fuel pollution harming including CO2e and particulate matter.  Note that in this graph, a high-payback ratio is a good thing and wind energy has the best payback of energy in both of its incarnations.
Cross-generation Wildlife Impacts
These types of analyses are focussed on downstream impacts of different forms of generation on wildlife and as such are directed assessments of environmental harm.
The most robust are two multi-energy source studies of wildlife mortality, including Comparison Of Reported Effects And Risks To Vertebrate Wildlife From Six Electricity Generation Types In The New York/New England Region, prepared for the New York State Energy Research And Development Authority in 2009, from which this table is reproduced. This is once again a full-lifecycle perspective on energy generation forms. (My perspective is that natural gas receives a better rating on the Power Generation column than it deserves as while it is much better than coal, it still generates 50 times the CO2e per KWh as wind energy or nuclear, but this study is solid otherwise.)
There are certainly other studies and mechanisms, but these are the studies and reports that I’ve run across that strive for apples-to-apples comparisons.
 ACLCA – Home
 LCA Harmonization
 Air sampling reveals high emissions from gas field