Measurements of tropical bamboo water use

Bamboos are grasses (Poaceae) that are widespread in tropical and subtropical regions. As monocots, woody bamboos have thin-wall hollow culms with abundant parenchyma and the culms are connected by rhizomes. These properties may imply that bamboos have specifc water use characteristics. For example, the abundant parenchyma in culms provides a water reservoir for internal water storage, which may influence the response of water use to environmental factors; the connected rhizome provides passages for water transfer between bamboo culms. However, only a few water-use related studies have been conducted with thermal dissipation probe (TDP) and deuterium (D2O) tracing method until now. Therefore, in the present study, we aimed at exploring water use characteristics of four tropical bamboo species (Bambusa vulgaris, Dendrocalamus asper, Gigantochloa atroviolacea and Gigantochloa apus) with TDP and deuterium tracing method. At the meantime, to accurately measure bamboo water use, we also tested and calibrated TDP on the studied bamboo species, and assessed the influence of bamboo water content on TDP. Additionally, as a less used water use measurement method, potential error sources of deuterium tracing method were explored when applying it on bamboos to explore water use, storage and transfer. As the frst step, the TDP method was calibrated with the stem heat balance method (SHB) and gravimetric readings in B. vulgaris culms in a pot experiment. In this experiment, it was confrmed that the stem heat balance method is well suited for bamboos but that thermal dissipation probes need to be calibrated. In situ, TDP was further calibrated with SHB for the four bamboo species and species-specifc parameters for such calibration formulas were derived. With the calibrated TDP, we monitored sap flux density (Js) of the four bamboo species along with three tropical tree species (Gmelina arborea, Shorea leprosula and Hevea brasiliensis) during a dry and a wet period. During feld monitoring, we found that some bamboo species reached high maximal Js. Across bamboo species, maximal Js increased with decreasing culm diameter. In the diurnal course, Js in bamboos peaked much earlier than radiation andiv vapor pressure defcit (VPD), and also much earlier than Js in trees. There was a pronounced hysteresis between Js and VPD in bamboos, which was less pronounced in trees. Three of the four bamboo species showed reduced Js at high VPD values during the dry period, which was associated with a decrease in soil moisture content. Possible roles of internal water storage, root pressure and stomatal sensitivity are discussed. Bamboos and other plants may substantially rely on stem water storage for transpiration. Fluctuations in wood water content (θwood) may lead to errors when estimating transpiration based on Js measurements with the widely used TDP method. To test the effects of θwood on Js, we conducted a culm dehydration experiment, monitored bamboos with TDP, and implemented a steady-state thermal model. Central to the calculation of Js, and thus a major potential source of error, is the maximal temperature difference between probes (∆Tmax) which are usually assumed to happen under "zero sap flow" condition. In the culm dehydration experiment, θwood was found to be highly negatively correlated with ∆Tmax. In the long-term feld monitoring, soil moisture content was also negatively correlated with daily ∆Tmax, indicating changes in θwood and a seasonal decrease in stem water storage. The steady-state model reproduced the θwood to ∆Tmax relationship of the dehydration experiment and underlined a considerable sensitivity of Js estimates to θwood. Fluctuations in θwood may lead to substantial underestimation of Js, and subsequently of transpiration, in commonly applied estimation schemes. However, our model results suggest that such underestimation can be quantifed and subsequently corrected when key wood properties are known. Our study gives insights into the relationship between θwood and TDP-derived Js and examines potential estimation biases. Deuterium tracing method was also applied on bamboos to explore water use, storage and transfer, and its potential error sources were explored. The deuterium-derived sap velocities correlated with the TDP-derived velocities in two of the three species. In all species, the deuterium residence time in bamboo culms was little influenced by the contribution of the culm water storage to transpiration as estimated by TDPs at different culm heights. Potential reasons include a small water storage volume in the culms, a low estimated contribution of the storage to transpiration and high sap flux densities. Daily culm water use rates estimated by the deuterium and the TDP approaches correlated linearly (R2 = 0.9) but were by 70% in the deuterium estimates. After the experiment, culms were cut down and analyzed for residual deuterium, but concentrations were low which indicates that retention did not play av major role in causing errors of the deuterium tracing approach. In culms neighboring the deuterium labeled culms of B. vulgaris and G. apus, elevated deuterium concentrations were detected indicating water transfer between culms. Based on the differences in daily water use on labeled culms and the enhanced deuterium concentrations in neighboring culms of these two species, we inferred that fve neighboring culms might receive water from the labeled culms. On contrast, in culms neighboring labeled D. asper culms, only slightly elevated deuterium concentrations were observed which implies a limited role by water transfer. However, incomplete mixing as indicated by high variation among three TDP sensors at a given height may be of particular importance for deuterium tracing in D. asper. In conclusion, species-specifc differences among big clumpy bamboos are indicated and the deuterium tracing points to water transfer among culms.