Methods
Leaf traits
Leaf trait data were compiled from both published and unpublished sources. Adata set was
considered suitable provided it contained data for at least two of the leaf traits for at least
four co-occurring species. Highly artificial vegetation types such as forestry plantations
and crop fields were not included. Only site-based data sets were used so that we could
reasonably attach climate data. The total data set (Supplementary Information)
represented 175 sites and contained 2,548 species–site combinations, consisting of 2,021
species, with 350 occurring at more than one site.
Climate
Climate data were taken from (1) the sites themselves, where known; (2) the nearest
weather stations, with temperature data scaled where necessary by an altitudinal lapse rate
of 0.6 8C per 100m(ref. 46); (3) a global 0.5 £ 0.58 data set ofMAT, MAR, vapour pressure
and irradiance47; (4) a global 0.5 £ 0.58 data set of Penman–Monteith PET calculated for the period 1987–1988 (ref. 48). MAT and rainfall data from the global data set agreed
closely with known or nearby-station data (scaled for altitude where necessary), giving us
confidence that the vapour pressure and irradiance data were reliable also. Irradiance data
were not adjusted for altitude because increases in cloud cover with elevation tend to offset
the increase in radiation that would be observed in clear air, although this may not be true
for high-elevation sites in arid regions46,49. For high-elevation sites, PET was estimated
from a regression equation fitted to all other sites, which considered PET as a function of
MAT and annual rainfall (r 2 ¼ 0.71). Vapour pressure was scaled for high-elevation sites
using an empirical formula expressing the exponential decrease of vapour pressure with
altitude49. Monthly mean VPD was estimated as the difference between the saturation
vapour pressure of air (at the monthly mean temperature) and vapour pressure taken from
the global data set. Saturation vapour pressure was calculated using the Tetens formula50.
Climate variables were averaged or summed (rainfall) across all months of the year, and for
those months with mean temperature over 4.99 8C, giving an estimate of climate during
the growth season. Results using yearly and growth-season climate indices differed little, so
we report results relating to yearly climate averages only.
MethodsLeaf traitsLeaf trait data were compiled from both published and unpublished sources. Adata set wasconsidered suitable provided it contained data for at least two of the leaf traits for at leastfour co-occurring species. Highly artificial vegetation types such as forestry plantationsand crop fields were not included. Only site-based data sets were used so that we couldreasonably attach climate data. The total data set (Supplementary Information)represented 175 sites and contained 2,548 species–site combinations, consisting of 2,021species, with 350 occurring at more than one site.ClimateClimate data were taken from (1) the sites themselves, where known; (2) the nearestweather stations, with temperature data scaled where necessary by an altitudinal lapse rateof 0.6 8C per 100m(ref. 46); (3) a global 0.5 £ 0.58 data set ofMAT, MAR, vapour pressureand irradiance47; (4) a global 0.5 £ 0.58 data set of Penman–Monteith PET calculated for the period 1987–1988 (ref. 48). MAT and rainfall data from the global data set agreedclosely with known or nearby-station data (scaled for altitude where necessary), giving usconfidence that the vapour pressure and irradiance data were reliable also. Irradiance datawere not adjusted for altitude because increases in cloud cover with elevation tend to offsetthe increase in radiation that would be observed in clear air, although this may not be truefor high-elevation sites in arid regions46,49. For high-elevation sites, PET was estimatedfrom a regression equation fitted to all other sites, which considered PET as a function ofMAT and annual rainfall (r 2 ¼ 0.71). Vapour pressure was scaled for high-elevation sitesusing an empirical formula expressing the exponential decrease of vapour pressure withaltitude49. Monthly mean VPD was estimated as the difference between the saturationvapour pressure of air (at the monthly mean temperature) and vapour pressure taken fromthe global data set. Saturation vapour pressure was calculated using the Tetens formula50.Climate variables were averaged or summed (rainfall) across all months of the year, and forthose months with mean temperature over 4.99 8C, giving an estimate of climate duringthe growth season. Results using yearly and growth-season climate indices differed little, sowe report results relating to yearly climate averages only.
การแปล กรุณารอสักครู่..
