2While climate change continues to present a major threat to global biodiversity and 2 3 ecosystems, most research on climate change impacts do not have the resolution to detect 2 4 changes in species abundance and are often limited to temperate ecosystems. This limits our 2 5understanding of global responses in species abundance-a determinant of ecosystem 2 6 function and services-to climate change including in the highly-biodiverse tropics. We 2 7 address this knowledge gap by quantifying abundance responses to climate change in 2 8 waterbirds, an indicator taxon of wetland biodiversity, at 6,822 sites between -55° and 64°.
9Using 1,303,651 count records since 1990 of 390 species, we show that with temperature 3 0 increase, the abundance of species and populations decreased at lower latitudes, particularly 3 1 in the tropics, but increased at higher latitudes. These contrasting responses to temperature 3 2 increase according to latitude indicate potential global-scale poleward shifts of species 3 3 abundance under climate change, providing empirical support for predictions by earlier 3 4 studies. The negative responses to temperature increase in tropical species and populations 3 5are of conservation concern, as they are often also threatened by other anthropogenic factors.
6Our results suggest that existing biases in studies towards temperate regions could 3 7 underestimate the impact of climate change on waterbirds and other species.3 8 3 9 4 1 urgent need to understand how species respond to changing climates globally. A wide range 4 2 of species have already shown responses to climate change, such as changes in geographical 4 3 3 range 1 , phenology 2 and abundance 3 . However, the rate and direction of these responses vary 4 4 greatly among species and locations 1,2,4 . As climate-driven changes in biodiversity are 4 5 expected to affect ecosystem functioning, human well-being, and the dynamics of climate 4 6 change itself 5 , understanding how species' responses to climate change may vary globally 4 7 could provide crucial evidence for a more effective allocation of limited resources on a 4 8 priority basis for the conservation of species and ecosystems threatened the most by climate 4 9 change, and for assessing how climate-driven changes in biodiversity may affect human 5 0 societies. 5 1Existing gaps in the geographical coverage of available evidence seriously limit our 5 2 understanding of species' responses to climate change and its variations across the globe 6 .
3Earlier global reviews of species' responses to climate change have rarely incorporated 5 4 species and studies in the tropics, due to the lack of relevant information 7 . Such geographical 5 5 biases are even more prominent in studies investigating responses in species abundance 8,9 , 5 6 which is a major determinant of species extinction risk 10 , ecosystem function and services 11 .
7Research on abundance responses to climate change to date has largely been conducted in 5 8 Europe, North America and the Arctic 3,12-14 , with a rec...