Deciduous and Evergreen Trees Rely on Deep Water Throughout the Year in a Subtropical Seasonal Forest

In subtropical and tropical seasonal forests, trees have adapted to low shallow soil water availability during the dry season by modifying root density, rooting depth, and leaf phenology. Here we test the well known hypothesis that water uptake in deciduous trees is restricted to the shallow soil layer, which prevents them from sustaining transpiring leaves during the dry season. Evergreens, on the other hand, access perennially available deep water sources, allowing them to maintain their transpiring leaves during the dry season. To determine where in the soil profile deciduous and evergreen trees take up water, we used stable isotope analysis to measure water source use of two deciduous and three evergreen species for a period of 13 months. In addition, to test the possibility that leaflessness could alter the isotopic composition of stem water, we measured the isotopic variation in stem water caused by artificial defoliation of an evergreen species. Deciduous and evergreen trees took up water from the same depths in both the wet and dry seasons. Deciduous and evergreen trees used approximately 51% deep water (50-150cm) throughout the year, while soil from 0-20cm was the least important water source with 24 and 6% of water uptake for wet and dry seasons, respectively. Low use of shallow water (0-20cm) in the wet season was due to inconstant water availability. Though the top 20cm of soil is the location of most nutrients, the soil’s limited water availability requires plants to have access to a more reliable deep water source to meet both their dry and wet season transpirational demands. This apparent spatial uncoupling in water and nutrient uptake denotes separate resource allocation for nutrient and water acquisition. Deciduous trees showed isotopic enrichment of stem water compared to evergreen plants only during the period that deciduous trees were leafless. We explain this as isotopic enrichment of fixed pool of stem water by evaporation as our defoliation experiment demonstrated. The δ18O and δ2H values of stem water from defoliated Q. virginiana branches were significantly higher than those of stem water from branches with leaves on the same tree (P < 0.01 and 0.001, respectively) and remained higher until the defoliated branches grew new leaves. Low sap flow in stems due to leaflessness or minimal transpiration can increase δ18O and δ2H values of stem water through evaporative enrichment. Using stem water to measure water source use in this situation would be misleading because it would infer usage of a water source more isotopically enriched than is actually the case.