news and update ISSN 1948-6596 commentary Ecological complexity at the forest–grassland transition revealed by lake sediment records Understanding how ecological processes play out over time scales relevant to long-lived species such as trees, and through periods of significant climate change, remains a major hurdle in trans- lating climate-change projections into biological impacts. Over the past few decades, paleoecolo- gists have embraced the prospect of using sepa- rate lines of evidence of climate, vegetation, and disturbance from sediment records to assess bio- geographic responses to climate change. Well- known complexities in responses to climate change include positive feedbacks within ecosys- tems, lag times, and the possibility of multiple po- tential states. This requires paleoecologists to have a robust understanding of past climate, and thus they have been thrust into the field of paleo- climatology. As there are multiple approaches to reconstructing past climate, researchers are often faced with the task of reconciling their climatic reconstructions with a myriad of existing paleocli- mate studies. In a new paper (Shuman et al. 2009), a cli- mate reconstruction based on fluctuating lake lev- els found evidence of a severe drought concurrent with a non-intuitive vegetation response. Shuman’s group studied the paleoclimatic context for the development of the Big Woods, Minessota, a large patch of mesic-adapted forest trees that occurred at the forest–grassland transition at the time of pre-European settlement but is now largely converted to agriculture and suburbia. Earlier paleoecological studies had determined that open woodland and grassland began being replaced by Big Woods vegetation at ca. 1300 AD, concurrent with a decrease in fire and, assumed but never addressed explicitly, increased moisture that favored tree establishment. The new study, however, claims that lake levels, and thus regional moisture, were lower during the Big Woods devel- opment than at any other time during the past 1800 years. The authors also show, using age dis- tributions of trees, that a comparable drought during the 1930’s was not severe enough to pre- clude establishment of mesic tree species such as sugar maple. The “ecological surprise” from these findings is that the local-scale controls of the for- est–grassland transition may operate, at times, in the opposite direction from the regional moisture gradient that explains the broad-scale pattern of the forest-to-grassland transition. The mechanisms underlying the Big Woods establishment, if indeed it happened during drought, require rethinking. Shuman et al. (2009) invoke a previously published hypothesis (Umbanhowar 2004) that fire may have declined because of drought-caused fuel discontinuity, and thus increasing fire-free intervals permitting re- cruitment of Big Woods tree species. This hy- pothesis fits broadly with the assertion that over large expanses of the planet where productivity is marginally sufficient for forest, the observed vege- tation is largely determined by fire and herbivores (Bond & Keeley 2005). The primary role of distur- bance over climate in these areas leaves wiggle- room for complexities including processes that reverse the climate–vegetation relationship at small spatial scales. One of these complexities not discussed with respect to the Big Woods is the effect of fluctuating populations of grazers, such as bison. The influence of grazer populations on the forest–grassland transition may be as great or greater than other factors, and because pa- leoecologists rarely discuss grazing impacts, they represent an “elephant in the room” in the inter- pretation of these records (Craine & McLauchlan 2004). For example, increased grazing pressure during drought may have been required to reduce fuel continuity and fire extent. Of course, devel- oping sediment proxies of grazer populations, such as from spores of the dung-fungus Sporormiella, would further the discussion of these dynamics (e.g., Gill et al. 2009; see also the update by Koch in this issue). The search for concordance among paleocli- mate records is a difficult task because of uncer- tainties in sediment chronologies, in the climate © 2010 the authors; journal compilation © 2010 The International Biogeography Society — frontiers of biogeography 2.1, 2010