Pinus caribaea Plantation Forest — Fiji

Introduction

This chapter presents the results of running VAMPS on a dataset concerning a Pinus caribaea plantation forest located on Viti Levu, Fiji [26]. A 61-day period (2 July — 2 September) from the Tulasewa site was modelled using meteorological data with 30-minute resolution.

The soils at this site have a bulk density ranging from 0.97 at the top to 1.07 g/cm at 1.2 m below the surface. Saturated hydraulic conductivity was determined for three soil layers from 29 core samples (Table 1). [26] determined Van Genuchten parameters for the topsoil (0–30 cm) and the subsoil (30–150 cm) using 12 and 17 samples respectively. These values were used in the soil section of VAMPS without any modification.

Transpiration was modelled using the Penman–Monteith combination equation while interception was modelled using an adapted version of the Gash model (described in the VAMPS model documentation). Soil evaporation was also modelled using the Penman–Monteith combination equation. It was assumed that net radiation at the forest floor was 3.5 % of that at the top of the canopy.

Measured values of moisture content (from a capacitance probe) were used to verify model results.

Soil parameters — Tulasewa site

Van Genuchten parameters used as VAMPS input (after [26]):

Soil layer	Porosity	Alpha [1/cm]	\(K_{sat}\) [cm/day]	n	Depth [cm]
1	0.60	0.061	1800	1.098	0 – 30
2	0.61	0.042	380	1.094	30 – 75
3	0.61	0.042	0.03	1.094	70 – 150

Results

The water balance summary for the modelled period is shown below:

Parameter	Value [cm]
Initial water volume of profile	71.765
Saturated water volume of profile	98.397
Total precipitation	25.767
Total transpiration	16.753
Total interception	5.261
Total soil evaporation	1.522
Total root extraction	10.693
Total drainage	−0.008
Total surface runoff	0.000
Initial storage	71.765
Final storage	80.197
Change in storage	−8.432
Percent mass-balance error	−0.039 %

The modelled interception and transpiration values are similar to those presented by [26]. Canopy resistance (\(R_s\)) was calculated using the relations derived by [26]. Calculated potential transpiration could not be maintained over the period due to modelled water stress; calculated root extraction (actual modelled transpiration) was 5.5 cm lower. This may be caused by:

The regression equations used to estimate \(R_s\) were obtained for a wetter period in which the trees did not suffer water stress.
The water content versus suction-head functions are not very accurate in the drier regions and may overestimate the suction head.

The average moisture content is generally modelled with good result. The upper layers show low moisture contents (theta ≤ 0.3); the pF corresponding to a moisture content of 0.3 is already beyond 4.2.

Example input files

The example input files for this case study are:

examples/fiji/fiji.inp — original run using Van Genuchten parameters.
examples/fiji/fiji_bandC.inp — equivalent run using Brooks–Corey parameters derived from the Van Genuchten values (λ ≈ n − 1, \(h_b \approx -1/\alpha\)). This file demonstrates method 6 and can be used to compare the two retention curve approaches.

The input file used for the Fiji study (abbreviated):

[vamps]
iniinmem=1

[run]
outputfile = run5.hh.out

[determine]
canopy = 1
soilmoisture = 1

[time]
steps = 2930

[ts]
precipitation = ../ninp/precip.prn
netrad        = ../ninp/rnet.prn
rhumid        = ../ninp/rh.prn
temp          = ../ninp/newt.prn
windspeed     = ../ninp/wind.prn

[interception]
method  = 0
E_avg/R = 0.147
p_f     = 0.6
p_tr    = 0.017
S       = 0.08
St      = 0.0062

[canopy]
transpiration  = 2
Rnet_absorb    = 0.975
layers         = 1
z              = 12.7
z_0            = 1.5
d              = 7.0

[roots]
swsink  = 0
swhypr  = 0
swupfu  = 0
depth   = 120.0
hlim1   = -5.0
hlim2u  = -50.0
hlim2l  = -50.0
hlim3h  = -800.0
hlim3l  = -1000.0
hlim3   = -1800.0
hlim4   = -12000.0

[soil]
layers        = 77
bottom        = 6
initprof      = 0
theta_initial = 0.200000 0.210000 ...  (77 values)

[layer_0]
description    = Tulasewa top layer
thickness      = 2.5
method         = 1
thetas         = 0.6
theta_residual = 0.08
alpha          = 0.061
n              = 1.098
l              = 0.5
ksat           = 1800

[layer_14]
description    = Tulasewa 30-75 cm layer
thickness      = 2.0
thetas         = 0.64
alpha          = 0.042
n              = 1.094
l              = 0.5
ksat           = 380.0

[layer_36]
description    = Tulasewa deep layer > 75 cm
thickness      = 2.0
thetas         = 0.6
ksat           = 3.0

Running with Python

The same run can be driven entirely from Python using vampspy:

from vampspy.model import Model

m = Model.from_file('examples/fiji/fiji.inp', vampslib='share')
result = m.run()
print(f"Total precipitation: {result['cumprec'][-1]:.3f} cm")
print(f"Final soil storage:  {result['volact'][-1]:.3f} cm")
print(f"Theta array shape:   {result['theta'].shape}")   # (61, 77)

For the Brooks–Corey comparison:

m_bc = Model.from_file('examples/fiji/fiji_bandC.inp', vampslib='share')
result_bc = m_bc.run()