Honestly that just makes me question how precise BRef's WAR metrics can really be, especially given how hard it is to quantify defense. The Fangraphs WAR is probably slightly more compelling for an individual award. And I can't say I expect voters to look at park effect very closely. At the end of the day they were three strong candidates and we should probably just be glad that the Nats didn't win the division or Scherzer's 18 wins and 300 strikeouts surely would have taken it.
Baseball-Reference loves Nola, to a large degree, because it's giving him a huge adjustment for the Phillies' bad defense -- 0.64 runs per nine innings. B-R estimates that the Phillies' defense cost Nola about 15 runs, which would drop his runs allowed from 57 to 42. Nola pitched in a neutral park while deGrom pitched in a pitchers' park, so that also increases the gap between the two. Maybe the Phillies' defense did hurt Nola; on the other hand, his .254 BABIP tied for the fourth-lowest among starters, so balls in play were turned into outs a large percentage of the time.
FanGraphs, meanwhile, sees that deGrom had a higher strikeout rate and lower walk and home run rates and deemed that he had the much superior season. Their calculations do not factor in Nola's hit rate or left-on-base rate or the fact that allowed fewer runs than you would expect because he held batters to a .129 average with runners in scoring position.