Global Rainfall Partitioning by Woody Plants in Drylands: General Patterns, Empirical Models and Invasion Processes.

MAGLIANO, PN; WHITWORTH-HULSE, JI; BALDI, G; CID, FD; LEPORATI, J; AGUIAR, S; ZEBALLOS, S; VAN STAN, JT; JOBBÁGY, EG

Nueva Orleans

Congreso; AGU Fall Meeting; 2021

Asociación Americana de Geo-Física

Rainfall partitioning by plant canopies can play key roles in terrestrial ecosystems by altering the amount, timing and patterns of water receipt to soils. Here, we developed three global meta-analyses that explored the role of plant morphological attributes (life-form, bark roughness, leaf type, leaf phenology) on rainfall partitioning into interception (I), throughfall (T) and stemflow (S) in drylands and worldwide invaded regions. First, we synthesized rainfall partitioning along a rainfall gradient (46 papers; 145 - 805 mm/y, 68 woody plant species). Second, we developed models to predict the event-scale response to rainfall partitioning (40 papers; 1,934 rainfall events modeled). Third, we compared the rainfall partitioning fluxes and their plant morphological controls for co-occurring invasive and native woody plant species (100 papers; 47 plant species comparisons, 12 regions). We found, first, I, T and S accounted for 24.0, 69.8 and 6.2% of total rainfall, respectively. Along the increasing rainfall gradient, I showed a tendency to decrease from 27.1 to 18.9%, T increased from 61.4 to 81.2% and S decreased from 10.0 to 1.6%. Shrubs presented more S than trees and were more abundant towards the dry edge of the rainfall gradient. Second, twelve linear models explained significant variability across all synthesized observations. S was best predicted by models that consider four canopy factors displaying slopes from 0.024 to 0.066 (mm of S per mm of rainfall). I was best predicted by leaf morphology (slopes from 0.162 to 0.171) whereas T best predictor was rainfall event size (slope = 0.822). Shrubs and smooth bark species present higher responses to S than tress and rough bark species, respectively. Third, invasive species presented higher S than native species for both drylands and humid areas, and higher T in drylands, but less in humid areas. Rough-barked species constrained rainfall inputs by promoting higher losses due to I, while smooth-barked species with broad leaves enhanced the amount of rainwater reaching the soil by maximizing S. Our findings highlighted the ecohydrological key role of vegetation life form and rainfall inputs affecting the amount of water entering into the soil, and suggested that specific morphological attributes of invasive species determine higher localized water inputs.