Abstract: From April to September in 2008, the stem sap flow velocity (Js) of se
veral common tree species (Ginkgo biloba, Aesculus chinensis, Magnolia denudata,
 Robinia pseudoacacia, Pinus tabulaeformis and Cedrus deodara) in Beijing was m
easured by thermal dissipation method. Cross-correlation analysis was used to e
stimate the time lag between the stem sap flow and the driving factors of canopy
 transpiration among the tree species. The Js of the six tree species was
significantly correlated with the total radiation (Rs) and vapor pressure
deficit (D), and the Js was lagged behind Rs but ahead of D.
 The maximum correlation coefficient of Js with Rs (074〖KG-*2〗-
〖KG-*7〗093) was often higher than that of Js with D (057〖KG-*2
〗-〖KG-*7〗079), indicating that the diurnal Js was more dependent on
Rs than on D. The sampled tree species except P. tabulaeformis had a
shorter time lag of Js with Rs (10〖KG-*2〗-〖KG-*7〗70 min)
than with D (47〖KG-*2〗-〖KG-*7〗130 min), and there existed significant di
fferences among R. pseudoacacia, P. tabulaeformis, and C. deodara. The t
ime lag between the Js and the driving factors of canopy transpiration was
 mainly correlated with the tree features (DBH, tree height, canopy area, and sa
pwood area) and the nocturnal water recharge, regardless of tree species.

Key words: urban tree species, thermal dissipation method, stem sap flow, time lag, nitrous oxide, localized emission factor, manure, spatial pattern, uncertainty estimation