Abstract: A satisfactory parameterization of root distribution, a key parameter for simulating root water uptake process, is important in improving simulation performance of land surface process models (LSMs) for simulating water and heat flux exchanges. Using observation data of maize root biomass in different growing periods in previous studies,
simulation performances  of three root distribution parameterization methods which established by Schenk, Zeng and Jackson and applied to the mainstream LSMs (defined as M1, M2 and M3, respectively)  were compared. The results showed that M1 could reproduce accurately root distribution profiles in every growing period while M2 and M3 had similar performances, with relatively high simulation precision only after corn tasseling stage. By analyzing the dynamic pattern of the simulated two parameters
(d₅₀ and d₉₅, the depths in which 50% and 95% of all roots were located) in different growing periods, we found that d₉₅ increased with maize growth and its variability was the biggest at jointing and maturity and the smallest at tasseling stage. But, d₅₀ only increased with maize growth before flare opening stage and was almost invariable after tasseling stage. With further investigating, a biggest gap for d₅₀ and d₉₅ could be found between the observed value and the set value in LSMs before flare opening stage, which is one of the important reasons for errormaking in land surface process simulating.

Key words: natural evergreen broad-leaved forest, microbial biomass nitrogen, nitrogen deposition, microbial biomass carbon