Discussion
Effects of prairie dog disturbance on diversity and CWM meansDisturbance by prairie dogs has been shown to affect multiple vegetation parameters (Connell et al. 2019; Duchardt et al. 2021). The main objective of this study was to understand the effects prairie dogs have on the GPCA El Tokio grasslands, using functional and taxonomic diversity measures. Contrary to our expectations, prairie dog disturbance showed an effect mostly when it was moderated by grassland type or season. In fact, only cover had an independent effect that was slightly higher without prairie dog disturbance. These findings are in contrast to studies stating that prairie dogs have negative effects on cattle due to their disturbance activities, e.g. forage consumption (Vermiere et al. 2004; Derner et al. 2006). Our findings in fact add to the literature of how environmental variables play a stronger role than grazing and animal disturbances on vegetation (Török et al. 2018; Grinath et al. 2019; Jäschke et al. 2020). It is also important to note that there were no independent effects with at least moderate evidence of prairie dog disturbance on the CWM height and leaf area, traits that are usually associated with grazing pressure (Diaz et al. 2007). In fact, CWM leaf area was only dependent on season and there was no evidence that CWM height was affected. We found moderate to weak evidence that prairie dog disturbance did filter C4 cover, which was higher in sites without prairie dog disturbances. This can be explained by the fact that prairie dogs prefer to feed on grasses (Mellado et al. 2005). Most grasses present in the study area are C4, specifically the grasses with highest cover such as Muhlenbergia villiflora in almost all grassland types, Sporobolus cryptandrus in agricultural grasslands, Aristida pansa in calcareous grasslands andBouteloua dactyloides in mountain grasslands, and so a lower cover would be expected. However, it is important to mention that a recent study, covering a period of 72 years (Augustine et al. 2017), showed that some of these C4 species are being outcompeted by C3 species in the long term, and have a larger decline without grazing effects.
In addition, we found that functional diversity, but not species diversity, responded to the joint effects of grassland type and seasonality with prairie dog disturbance, confirming not only the need of including multiple environmental variables and their interactions to identify ecosystem complexity (Dainese et al. 2015), but also the importance of considering functional diversity to further understand the instances of these patterns (Cadotte et al. 2011). Prairie dog disturbance moderated FSpe in agricultural grasslands, possibly explained by the suppression of rapid growing species with extreme traits (eg. Salsola kali, Machaeranthera tanacetifolia, Kochia scoparia ) that have higher LA and height and are able to grow and dominate in agricultural grasslands without prairie dog disturbance. These species grow despite the lack of ideal water availability and soil conditions, because they benefit from the gain of resources due to nutrients from fertilization that remain after abandonment (Laliberte et al. 2012). Prairie dogs need short vegetation for predator avoidance (Hoogland, 1995), and their suppressing effect has been shown in previous literature (Ponce-Guevara et al. 2016; Hale et al. 2020). However, since the mechanism behind this suppression is unclear, further studies are needed to determine whether prairie dogs colonize agricultural grasslands before or after rapid-growing species have a chance to grow, or whether other mechanisms are at work (e.g. drought avoidance; Blumenthal et al. 2020). We further found no evidence of positive or negative effects on FEve and FDiv between grassland types, nor between any other of the measured conditions. This indicates that traits were mostly unchanged in their distribution and abundance between communities in the functional space volume. We can assume that the redundancy of traits is most likely increasing due to the restrictive environmental conditions that only well-adapted species can withstand (Villéger et al. 2008, Mouillot et al. 2013).2) Grassland types as major effect drivers
We found that grassland types explained most of the effects on plant functional and taxonomic diversity as well as CWM of traits. There was very strong evidence that mountain grasslands were positively affected in almost all measures, usually followed by arid, calcareous and agricultural grasslands, respectively. This can be explained by the tendency of mountain grasslands to have leptosol soils, highly variable slopes as well as higher elevation and lower atmospheric pressure, leading to higher precipitation and lower temperatures (Gommes, 2002; Anjos et al. 2015). These conditions are known to often cause an increase in plant species richness and cover (Speed et al. 2013; Buzhdygan et al. 2020). In addition, Pando-Moreno et al. (2013) found that many of the sites in mountain grasslands within GPCA El Tokio had a lower level of electrical conductivity and absence of gypsum, while sites that fall within calcareous, arid and agricultural grassland types had at least some percentage of gypsum in them. The presence of gypsum most likely acts as a habitat filter for CWM traits, as gypsum soils are known to limit plant life due to their chemical and physical properties which restrict plant growth (Escudero et al. 2015). Calcareous and mountain grasslands have similar filtering effects on perennial and C4 cover; the effect is most likely due to calcareous soils having lower gypsum levels. Although gypsum soils have been classified as calcareous soils and vice versa in many soil classification systems (Herrero, 2004), they have distinct effects on biodiversity, and it is even possible that a mix of both soil types could lead to higher biodiversity (Meyer et al. 1992; Luzuriaga et al. 2015). Arid grasslands in this study are also dominated by gypsum soils and have higher temperatures which, together with low precipitation, result in higher level of aridity which can act as a strong filter for most CWM traits (Vicente-Serrano et al. 2012; Munson et al. 2013). Similarly, the strong filtering effect of agricultural grasslands is most likely due to the lack of evolutionary history of vegetation, where surrounding landscape factors might play a greater role in species establishment due to continuous land use changes over many years (Gustavsson et al. 2007), paired with the restrictive conditions faced by vegetation on gypsum and calcareous soils (Meyer et al. 1992). Future studies, disentangling the climatic and edaphic effects of these grasslands types are needed to properly understand these patterns (Le Bagousse-Pinguet et al. 2017). On the other hand, the Inverse Simpson evenness showed an opposite result. It was higher for agricultural grasslands, which may be explained by the fact that the index assigns a higher evenness value to communities with an almost equal amount of rare and dominant species (Smith and Wilson, 1996). Hence, the higher the amount of rare species is, the lower is the Inverse Simpson evenness (Magurran, 2004).