Review: Henne et al., Increased burning in a warming climate reduces carbon uptake in the Greater Yellowstone Ecosystem despite productivity gains

This is a good, comprehensive look at the relevant issues here. I had a few comments about characterization of the work and in particular about its applicability to the rest of temperate North America. I think that claim is unsubstantiated by the analysis and I point out why below. One last issue that the readership at JE might think of is the role of invasives in changing surface fuels. I think it is unlikely to be a problem sufficient to change the outcome of the analysis in the GYE, especially since surface fuel communities prone to invasion are currently a smaller area of the system, but worth noting, perhaps near where insects are mentioned in the discussion caveats.

Line by line comments:

Line 28: “trigger carbon release”; This implies that it wasn’t already happening – I don’t dispute that it can increase abruptly, but carbon release via respiration is an ongoing process, so maybe this isn’t the best word choice.

Line 48-50: “Temperate mountain forests in western North America can maintain a carbon sink through the mid-century in a warming climate but continued warming later in the century may flip them to a carbon source.” The veracity of that synthesis depends entirely on whether the area examined in this paper (the GYE) is a surrogate for all temperate mountain forests in western North America. I would argue that it is not – plenty of other temperate forests have (1) different expected climate changes and (2) different climate-fire relationships and therefore the conclusion that what happens in Yellowstone is a proxy for the rest of the temperate continent is not legitimate. The authors should be careful of such sweeping claims.

Line 122-123: “…and the elimination of vast areas of forest cover (Littell et al., 2018; Westerling et al., 2011)” Littell et al. 2018 does not model vegetation and thus does project vegetation responses. It is logical to conclude that the increase in area burned would result in large areas of at least temporarily biome transition, but Littell et al. 2018 make no claims about implied elimination of forest cover. The citation here should be moved to after area burned, e.g., “…..future climate projections indicate major increases in area burned by the middle of the century (Littell et al. 2018), with the potential transformation of GYE forests in to a net C source, and the elimination of vast areas of forest cover (Westerling et al. 2011).”

Line 191-192: “…Thornthwaite- type approach.” There are limitations to this method compared to a Penman-Monteith that will be important to document briefly in the discussion (currently there is acknowledgment in the SM). In particular, Thornthwaite PET increases universally with temperature, whereas a Penman can produce the more realistic nonlinear response associated with relative humidity responses. The authors acknowledge this in the next sentence, but they are correct only for “defining the OBERVED landscape-scale variation….”; the FUTURE variation may or may not be appropriately simulated and this needs to be addressed as a caveat later in the paper. A Thornthwaite approximation shouldn’t be considered a complete water balance model because it doesn’t adequately simulate the interaction between the water and energy balance components. If you force a hydrologic model to solve for the energy balance and water balance simultaneously, you get a different result. In the GYE, it may not matter practically because of the continental climate, but in other temperate forests (like those you extrapolated to in the abstract) it matters a lot, and summer humidity changes and the resulting temperature response could actually increase much less than a Thornthwaite would lead you to believe.

Line 335: “….under the RCP 8.5 emissions scenario.” The authors need to justify this choice. Current thinking in climate impacts is that it is now unlikely for emissions to meet or exceed RCP 8.5 because it is based on a no-policy family of scenarios that use considerably more coal and possibly oil than current trajectories indicate. Management relevant, decision appropriate information should include a range of emissions scenarios (e.g., consistent with the likely range of emissions, say from RCP 4.5 or RCP 6.0 to RCP 8.5). It’s also true that until the 2060s or later, it doesn’t matter much. If the range of GCM climates is sufficiently broad, the authors could also argue that the impacts from warmer models under RCP 4.5 overlaps with the lower end of RCP 8.5, but that isn’t shown here. A caveat in the discussion is therefore also required.

Line 624-627: “The range of drying projected in our future climate scenarios encapsulates climatic changes expected for the forested regions of western North America, particularly north of 40° N (Christensen et al., 2013; Joyce, Talbert, Sharp, & Stevenson, 2018; Maloney et al., 2014).” The northern Great Basin as well as the Cascades and Coast Ranges may be exceptions to this for two reasons. First, in the Great Basin, the changes in convective precipitation and any related lightning activity are different from the GYE and also subject to considerable variation across GCMs, both in terms of historical skill and future projected changes. Second, in the Cascades and coast ranges, which have a modified Mediterranean climate, the change in summer precipitation is still uncertain, but the proportional change in summer aridity might even be greater in some places than in the central Rockies. The authors need to be more circumspect about the regional applicability of this range of simulations and outcomes.

Line 665: “Inset outbreaks…may increase in frequency in a warming climate….”. At first, yes, but as the climate becomes unfavorable for univoltinism, it may eventually decrease at a given elevation, while increasing at higher elevations. This is good to caveat, but it might be more important to suggest a complex response.

The authors have satisfied my previous comments.