Water use, storage and transfer in tropical bamboos

Bamboos (Poaceae, Bambuseae) are abundant in the natural vegetation of tropical and subtropical regions. They have been used by humans for millennia, and multiple additional bamboo usages have been developed in recent decades. From much better-studied water use characteristics in dicot tree species, the water use patterns of bamboos could potentially vary substantially due to several aspects. Bamboos are monocotyledonous species and lack secondary growth, and vascular conduits of bamboo xylem thus have to remain functional throughout the ontogeny of a bamboo culm. The culm is mostly hollow and high in parenchyma content, which might lead to a high potential for stem water storage. Additionally, culms are often connected via rhizomes, which offers pathways for a redistribution of water among connected culms. In this study, we addressed water circulation patterns in bamboos with particular respect to (1) water use rates, (2) the role of the stem water storage, and (3) water transfer among culms. Applied methods include thermal sap flux measurements (thermal dissipation probes, TDP; and stem heat balance, SHB), deuterium tracing and modeling. We conducted lab and field experiments, and monitored water use in the field. Field studies were mainly carried out in southern China and on Java, Indonesia. Four bamboos species were studied (Bambusa vulgaris, Dendrocalamus asper, Gigantochloa atroviolacea and Gigantochloa apus) in parallel to several tree species for comparison. To measure sap flux density of bamboos as precisely as possible, in the first step we calibrated the TDP method in the laboratory and the field. In both potted plant and field calibration experiments, we showed that the TDP method was suitable and reliable for sap flow measurements on bamboos after calibration. The potted plant experiment confirmed commonly reported underestimations of the TDP method as well as the accuracy of a second sap flux method, the stem heat balance technique. However, estimates with 1-cm-length TDP were significantly linearly correlated with those by the stem heat balance and a gravimetric method. From these linear relationships between estimates of the two methods, we subsequently derived species-specific calibration parameters for the TDP method for each of the four bamboo species in situ. The species-specific calibration parameters varied significantly among species, reflecting differences in wood thermal properties even though all four studied bamboo species are big, arborescent bamboos. To gain a better understanding of the effects of wood thermal property on TDP measurements and accuracy, we examined changes in TDP-derived sap flux density with changing stem water content in bamboo culms. In a three-step approach, we conducted a culm dehydration experiment in the laboratory, monitored bamboos with TDP in situ and lastly implemented a steady-state thermal model. One central assumption of the TDP method is a constant wood thermal conductivity for determining maximum temperature differences between heated and reference probes of the TDP sensors. However, due to dynamics in bamboo culm water content, these assumptions may not always hold. In the dehydration experiment and numerical modeling simulations, we showed that wood thermal conductivity decreases with decreasing culm water content, which led to increasing maximum temperature differences and consequently to an underestimation of sap flux density if the change in water content was ignored. Keeping other controlling variables constant, we found that underestimations became particularly apparent 1) for large decreases of water content from nighttime to daytime, 2) at relatively low sap flux density, and 3) for relatively larger nighttime water content when the ratio of decline to the daytime was kept constant (e.g., by one half). Based on these insights, we provide a novel logistic-regression correcting equation using three parameters related to nighttime and daytime water content. Bamboos were found to maintain relatively higher maximal sap flux density (21.6 – 70.5 g cm-2 h-1) compared to nearby tree species (10.5 - 23.3 g cm-2 h-1; Gmelina arborea, Shorea leprosula and Hevea brasiliensis). However, we did find evidence of diurnal hysteresis between sap flux density and transpirational driving variables (e.g., vapor pressure deficit and radiation), which indicated a decoupling from atmospheric processes and a rapid withdrawal of water from the stem during the morning hours. Observations with TDP at the top and bottom of selected bamboo culms manifested that stem water storage mechanisms supported only about 10% of daily transpiration of bamboos. Further, the relatively short water residence time in bamboo culms (5.5 – 6.3 days) as derived from deuterium tracing methods also implied a rapid exchange of water between conducting passways and storage compartments and thus a limited capacity for stem water storage in bamboos. Therefore, for the studied bamboo species, stem water storage mechanisms were shown to be of less importance than we had assumed considering the significant percentages of parenchyma tissue in bamboo culms. The prior study points to only small contributions of internal stem water storage to daily transpiration of bamboos. Nonetheless, the interconnected underground rhizome systems of bamboos can provide additional water transfer to cope with and balance water demand. The deuterium tracing applied in this study pointed to the existence of below-ground water transfer among culms within the same clumps. This mechanism provided significant transfer to newly sprouted culms. For newly sprouted culms, about 48% of daily water use may be supplied by established neighboring bamboo culms. When all established culms in a certain clump were removed, water use of newly sprouted culms could drop by about 80%. The dependency of newly sprouted culms from established culms largely disappeared after they started to produce leaves. However, after about four months since their emergence, newly sprouted culms kept active nighttime sap flow regardless of whether they were in leaves or not. To explore the mechanisms behind these somewhat unusual water use patterns, we focused on and discussed in detail potential driving forces as well as possible interactions with nutrient redistribution mechanisms. At last, as a case study in a tropical dry forest in Ecuador, stem water storage estimated with the deuterium tracing method was assessed on a stem-succulent tree species (Ceiba trichistandra) and two coexisting evergreen species (Capparis scabrida, Geoffroea spinosa) and two other deciduous species (Eriotheca ruizii, Erythrina velutina). Unlike the relatively small stems of bamboos, stem-succulent trees are characterized by large, ‘swollen’ trunks that can reach several meters in diameter. Large water storage ability of succulent stems was confirmed by the relatively longer water residence time (21.7 days) of a deciduous stem-succulent tree species (C. trichistandra) in contrast to the two evergreen species (on average 12.8 days; C. scabrida, G. spinosa) at the same sites. However, the water residence time of C. trichistandra (average diameter 76 cm) did not show significant differences to the two accompanying deciduous species (E. velutina and E. ruizii) with average diameters of 28 and 56 cm (water residence time 19 days). A literature review of woody species (n=33) in which water residence time was estimated with deuterium tracing suggested, that residence time increases with plant diameter. The bamboos from the previous study in Java as well as the dry forest trees in Ecuador fit well into this picture, with the bamboos at the low diameter/short residence time end and the stem succulent C. trichistandra at the big diameter/long residence time end. Additionally to its diameter-dependency, residence time was found to increase significantly with decreasing wood density. In conclusion, the studied bamboos reached comparably high maximal sap flux densities. Culm water storage did not contribute substantially to daily transpiration. The residence time of water in bamboos culms was at the lower end of all available studies. The interconnected underground rhizome systems of bamboos were used for water transfer in particular to newly sprouted, leafless culms. Further exploring such underlying eco-hydrological mechanisms in greater detail will be a challenge for the years and studies to come.