How do rubber (Hevea brasiliensis) plantations behave under seasonal water stress in northeastern Thailand and central Cambodia?

Abstract

a b s t r a c tDelineating the characteristics of biosphere-atmosphere exchange in rubber (Hevea brasiliensis Müll. Arg.) plantations, which are rapidly expanding throughout mainland Southeast Asia, is important to understanding the impacts of the land-use change on environmental processes. In attempt to shed new light on the impacts of conversion to rubber, we have conducted eddy flux measurements over a 3-year period in two rubber plantation sites: (1) Som Sanuk, located in northeastern Thailand; and (2) Cambodian Rubber Research Institute (CRRI), located in central Cambodia. Both sites have a distinct dry season. We used a combination of actual evapotranspiration (E T ) flux measurements and an inverted version of a simple 2-layer E T model for estimating the mean canopy stomatal conductance (g s ). The potential water balance (precipitation (P) − potential evaporation (E T POT )) for each season (i.e.revealed when and how the water use is controlled. In the seasons when actual water balance (P − E T ) was negative (DJF and MAM), the deficit was compensated with soil water from the previous season at depths of 0-2 m (Thailand site) and 0-3 m (Cambodia site). At both sites, the reference value of g s (g sref ) and the sensitivity of g s to atmospheric demand (m) appeared to be less in DJF and MAM than each in the other two 3-month periods (seasons). On average, in a whole year, m/g sref was less in Thailand («0.6) than in Cambodia (near 0.6 for part of the year), suggesting that there was less sufficient stomatal regulation at the Thailand site, where there might be little risk of water stress-induced hydraulic failure because of its higher annual rainfall amount. In comparison, at CRRI where annual P − E T POT was negative, there was stricter stomatal regulation, preventing excessive xylem cavitation.

Rubber tree transpiration in the lowlands of Sumatra

“…Climate in the mainland is highly seasonal, with clear annual cycles of precipitation, soil moisture, and VPD (Giambelluca et al, 2016;Kumagai et al, 2015), whereas in our study region, there was a lack of a pronounced dry season, soil moisture remained high, and VPD showed relatively low intra-annual variability. Climate in the mainland is highly seasonal, with clear annual cycles of precipitation, soil moisture, and VPD (Giambelluca et al, 2016;Kumagai et al, 2015), whereas in our study region, there was a lack of a pronounced dry season, soil moisture remained high, and VPD showed relatively low intra-annual variability.…”

“…We used micrometeorological data from a station located approximately 15 km from "Harapan region" and 60 km from "Bukit Duabelas region" (PT. To ensure comparability to other studies (e.g., Kumagai et al, 2015), we thus only adapted location-related constants, and default values were used for all other constants (e.g., Priestley-Taylor coefficient: 1.26). To ensure comparability to other studies (e.g., Kumagai et al, 2015), we thus only adapted location-related constants, and default values were used for all other constants (e.g., Priestley-Taylor coefficient: 1.26).…”

“…The studies in the Asian mainland were at least partially conducted in larger plantations (e.g., in an experimental farm in Cambodia; Giambelluca et al, 2016;Kumagai et al, 2015), where more intensive management, for example, with regard to optimal harvesting practices or the utilization of higher yielding rubber varieties, may induce higher productivity, which would likely go along with higher rubber tree water use. Relatively low water use rates of the mature small-holder plantations investigated in our study are in line with relatively low NPP (7.7 ± 1.1 Mg ha −1 year −1 ), which was for example, reported to be 53% and 34% lower than that of oil palm plantations and forests in the study region (Kotowska et al, 2015).…”

Mapping the Expansion of Boom Crops in Mainland Southeast Asia Using Dense Time Stacks of Landsat Data

“…All footprints except 126050 show a larger number of rubber change classes than all the other boom crop change classes combined. We assume that classifying rubber worked better because the selection of seasonal images (Figure 3) targeted the distinct phenology of the deciduous rubber [15]. Rubber therefore shows higher accuracies than the evergreen tree crop classes because rubber sample points overlapped less with other classes in the feature space, even if the number of sample points were unbalanced [56].…”

“…To capture the distinct phenology of different land-cover types and land-use activities, we selected Landsat images with acquisition dates from multiple seasons during the fifteen-year study period (Figure 2). Distinguishing rubber from forests, for example, requires capturing the timing of rubber's leaf defoliation (January, February) and the rapid re-greening phase (March, April) [15]. Deciduous forests, on the other hand, stay defoliated throughout the dry season and do not foliate as fast or as intensively as rubber [39].…”