Quantifying transpiration rates across stands and landscapes is a critical component for quantifying forest hydrology. Commonly, stand transpiration is quantified by installing sap flow probes in a subset of trees, deriving a relationship with an easily measured variable (such as DBH) and measuring DBH in all trees across a stand. Some studies have highlighted how scaling relationships can vary across species, forests types, or climates while others have found remarkable similarity, suggesting a “functional convergence” of tree functioning. The accuracy of these scaling relationships is not trivial, as small differences in the modeled relationship and variation in stand structure can result in large differences in stand transpiration. This is particularly challenging in tropical regions that consist of many competing land uses contributing to complex forest structure and confounding issues between tree age and size. In this study, we examined these scaling relationships across three different forest ages and one shade coffee plantation prevalent in the tropical montane region of Veracruz, Mexico. We quantified sap flow rates in forty trees across two regenerating forests (20 and 40 years), one mature forest, and one coffee plantation and examined how scaling relationships changed across species, stands (land cover types), and seasons.
We found remarkable convergence of the sap flow-DBH relationship within each stand. R-squared values within sites were greater than 0.93 across all sites in wet, dry, and winter fog seasons. When sap flow-DBH relationships were culled across stands, there were consistent patterns suggesting similar relationships across land cover types despite trees of various ages and species. Using the within-site relationships to quantify stand transpiration rates resulted in similar rates between the 40-yr regenerating forest and the mature forest. The shade-grown coffee and 20-yr regenerating forest sites had transpiration rates approximately 75% and 25% of the mature forest, respectively. Estimates of stand transpiration using the universal relationship varied by less than 5% from estimates using site-specific relationships and fell within natural ranges of variation across stands. Additionally, separating analysis periods into distinct seasons resulted in distinct relationships across seasons highlighting the importance of capturing seasonal variation. These results suggest a strong convergence in plant functioning while also highlighting the importance of capturing sources of annual variation. We further show that by taking these factors into account, relationships can be reliably extrapolated across sites, thereby allowing for reliable scaling of transpiration rates due to minimal variation relative to variation at larger scales.