Stream flow, ground water and climate change 
111 
reservoirs, the inclusion of which, being an important water source for the catchment, 
is crucial for the numerical surface water and groundwater simulations. 
Model setup 
Table 1 summarises the two scenario runs performed in this study. Scenario 1 serves 
to search for the steady state condition of the water balance in the system, which 
was modelled beginning with fully saturated conditions and run for many years with 
assumed zero rainfall and evapotranspiration (ET). The steady state water distribution 
resulting from Scenario 1 is then used as the initial condition for the real simulation in 
Scenario 2, which considers the real operational reservoir water levels and observed 
rainfall as the driving forces. For tropical swamp forests, evapotranspiration plays an 
important role in the entire water balance and hydrological cycle. Scenario 2 utilises 
a 2-layer water balance model to simulate the water loss from ET and the unsaturated 
zone storage. The 2-layer water balance model is a simplified water balance method 
which divides the unsaturated zone into a root zone and a zone below the root zone; 
ET can be extracted from the root zone, w hil e it does not occur in the zone below (Yan 
& Smith, 1994). 
Setting up the 2-layer water balance model essentially requires three inputs, i.e., 
the root depth (RD), the leaf area index (LAI) and the reference ET. The root depth is 
calculated based on a linear equation 
RD = 0.07624 x DBH + 0.11185 
where DBH is the diameter at breast height measured in 40 vegetation plots distributed 
across the Nee Soon catchment (Chong et al., 2018). The average value is used as the 
representative root depth in each plot; Thiessen polygon is then applied to interpolate 
the point values into the entire study area. 
The reference ET is the rate of ET with an unlimited amount of water from a 
reference surface - a hypothetical grass reference crop with specific characteristics 
(Allen et al., 1998). The reference ET data is obtained in this study from MOD 16 
Global Terrestrial Evapotranspiration Data Set (Mu et al., 2013). The MOD16 
project is part of the National Aeronautics and Space Administration (NASA)/Earth 
Observing System (EOS) project to estimate global terrestrial evapotranspiration from 
earth land surface by using satellite remote data. The MOD 16 dataset is derived from 
an improved ET estimation algorithm with inputs including the GMAO and MODIS 
land cover, LAI, FPAR and albedo data (Mu et al, 2011). Fig. 2 (a) shows samples of 
the MOD 16 reference ET over the Nee Soon freshwater swamp forest. The missing 
data accounts for about 1.2 km 2 within the Nee Soon catchment (less than one third 
of the catchment area), which are supplemented with the interpolated values from 
neighbouring cells. The 2012 reference ET ranges from 65 to 140 mm /month with an 
average of 100 mm/month. 
The Leaf Area Index is a dimensionless quantity that characterises plant 
canopies. It is defined as the one-sided green leaf area per unit ground surface area in 
