Climate change and primary production: Forty years in a bunchgrass prairie

Richness and diversity may in part have increased as the incidence of invasive species increased by 108%, as summer senescence intensified, which is consistent with observations that grassland invasives increase with drying conditions [86]. Increasing invasives had no effect on NBR ANPP (P < 0.28), and thereby, would have little effect on summer senescence measures, as > 50% of invasives were annual brome grasses, which provide very little biomass (cheatgrass, Bromus tectorum, accounted for less than 3% of the invasive bromes). Average plant drought resistance increased by 35.8% over our study, which is largely due to the increase in wheatgrasses and decrease in bluegrasses. Fire tolerance was unchanged over our study. Future of bunchgrass prairie Seasonal, rather than annual, changes in climate and their strong directional effect on the NBR vegetation, may be affecting the small areas of remaining bunchgrass prairie elsewhere. Climate change can be directional or part of a natural cycle. Using time series partial autocorrelation and Fourier analysis on the 109 year NBR climate record, one can observe 4–6 and 10–15 year cycles in precipitation and temperature, as expected, given cyclic oceanic drivers [87–89]. These cycles cannot account for the 40-year directional changes observed, but can only modify them [89,90]; rather an unidentified cycle with periodicity greater than twice our 40-year study period would be needed to produce directional changes, which seems unlikely. Therefore, a likely explanation for our observations is anthropogenic climate change [7,75,87,91–93].

The identified seasonal climate changes also may impact intermountain bunchgrass prairies in other ways, including increased invasive species impacts and susceptibility to fire. Invasives are dramatically increasing at other intermountain bunchgrass prairies [94]. Invasives do not influence NBR ANPP, because most are small annual plants that do not add much biomass (>50% are annual grasses). However, these annual plants contribute to increased summer senescence and thereby, may intensify the potential for fire. Historically, intermountain bunchgrass had low fire incidence compared to Great Plains grassland due to less frequent lightning [36,37], but higher summer temperatures spawn more lightning [95–97], and summer temperature is increasing in the bunchgrass. Therefore, greater ANPP and summer senescence in NBR bunchgrass provides more fuel for increased lightning to spawn fires than in the past. Furthermore, we found that plant species that are more resis- tant to fire are not increasing in abundance, further enhancing the potential for fire disturbance. What the intermountain bunchgrass will look like in the future given climate changes and associated disturbances is uncertain? There is very little difference between our study valley with bunchgrass prairie (Mission Valley, MT) and an adjacent valley (Little Bitterroot Valley, MT) (respectively, average annual precipitation of 40.5 vs. 37.5 cm and annual temperature of 7.4 vs. 7.5˚C) [98]. There is no evidence that the less productive cold desert steppe [99] will

replace the bunchgrass, as steppe is dominated by shrubs and we find that shrubs are declining (P < 0.004, N = 35) along with other dicots. This is counter to Harte et al. [14] reporting that woody vegetation is replacing Colorado grassland with observed and experimental warming. If the observed increase in NBR grass ANPP and relative abundance continues, then historic bunchgrass ecosystem may morph into a different, possibly previously unknown, type of grass- land. This uncertainty is consistent with claims that local and seasonal climate change may cre- ate very different ecosystems [8,13,23–25]. Conclusion Climate over the past 109 years in a bunchgrass ecosystem has annually warmed and dried as expected with anthropogenic climate change. Over the past 40 years, we found that changes in PLOS ONE Bunchgrass prairie and climate change PLOS ONE | https://doi.org/10.1371/journal.pone.0243496 December 23, 2020 15 / 21