Development and validation of Sawit, an oil palm growth and yield model
A study was undertaken to develop and validate Sawit, an oil palm growth and yield model that considers weather variables, planting densities and soil textures. Sawit consists of five components: meteorology, photosynthesis, energy balance, water balance, and crop growth. The meteorology componen...
Saved in:
| Main Author: | |
|---|---|
| Format: | Thesis |
| Language: | English English |
| Published: |
2023
|
| Subjects: | |
| Online Access: | http://psasir.upm.edu.my/id/eprint/112560/1/FP%202023%204%20-%20IR.pdf http://psasir.upm.edu.my/id/eprint/112560/ |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | A study was undertaken to develop and validate Sawit, an oil palm growth and
yield model that considers weather variables, planting densities and soil
textures. Sawit consists of five components: meteorology, photosynthesis,
energy balance, water balance, and crop growth. The meteorology component
was parameterized using daily weather data recorded in oil palm plantations.
The photosynthesis component was parameterized through leaf measurement
on young to mature palms, where photosynthetic parameters were measured.
Measurements of trunk and canopy heights, as well as the relationship between
oil palm leaf stomatal conductance with photosynthetically active radiation and
air vapour pressure deficit, were taken to parameterize the energy balance
component using the Shuttleworth-Wallace model. These measurements were
used to model evapotranspiration and leaf temperature. Water flow in the soil
was modelled following the 'tipping bucket' system using Darcy's law. The crop
growth component was parameterized based on the OPSIM model. Different
ages of oil palms were destructively sampled to determine the dry matter
partitioning and respiration coefficients. Sawit was validated by oil palm
planting-density experiments in six different sites: Rengam, Merlimau,
Kerayong, Sungai Buloh, Sabrang, and Seri Intan. The model showed good
agreement in predicting oil palm growth and yield parameters. However, the
accuracy of the simulation varied considerably between sites and parameters.
The yield simulation was considered sufficiently accurate for Merlimau and
Sungai Buloh, with the refined index of agreement (dr) equal to 0.76 and 0.74,
respectively, and the normalized mean absolute error (NMAE) equal to 0.19 and
0.20, respectively. However, the yield was simulated as less than satisfactory for
Rengam, Kerayong, Sabrang and Seri Intan. Their dr values were small (0.52-
0.56), but NMAE (0.13-0.25) were comparable to Merlimau and Sungai Buloh.
The simulation of vegetative dry matter production was good for Rengam (dr =
0.78, NMAE = 0.20), Kerayong (dr = 0.71, NMAE = 0.14) and Sabrang (dr = 0.76,
NMAE = 0.11), but satisfactory for Merlimau (dr = 0.62, NMAE = 0.23) and
Sungai Buloh (dr = 0.64, NMAE = 0.21). Total dry matter production was
simulated sufficiently accurate with dr ranging from 0.71-0.78 and NMAE
ranging from 0.07-0.17 across all sites and planting densities. Rachis (dr = 0.68-
0.86, NMAE = 0.11-0.19), fronds (dr = 0.65-0.83, NMAE = 0.11-0.20) and trunk (dr
= 0.69-0.82, NMAE = 0.24-0.39) were simulated more accurately than pinnae (dr
= 0.43-0.73, NMAE = 0.18-0.30) across all sites and planting densities except for
the simulation of trunk biomass in Sungai Buloh (dr = 0.38, NMAE = 0.77). The
leaf area index was simulated sufficiently accurate for Merlimau, Rengam, and
Seri Intan, with dr ranging from 0.78-0.84 and NMAE ranging from 0.03-0.16. In
contrast, the leaf area index was simulated as less than satisfactory for Kerayong,
Sungai Buloh, and Sabrang, with dr ranging from 0.57-0.62 and NMAE ranging
from 0.13-0.17. The simulation of trunk height was especially good for
Merlimau, Kerayong, Seri Intan and Sabrang, with large dr values (0.87-0.93) and
small NMAE values (0.06-0.12). However, trunk height was simulated
satisfactory for Rengam (dr = 0.67, NMAE = 0.39) and Sungai Buloh (dr = 0.69,
NMAE = 0.32). In addition, Sawit effectively simulated the impacts of El Niño
event on oil palm yield. It also accounted for the influence of soil textures,
rainfall, planting densities, and meteorological factors on water deficits.
However, the simulation errors increased with increasing planting density due
to insufficient characterization of microclimate conditions and plant water
uptake under dense oil palm canopies, and higher variability of measurements
for higher planting densities. In conclusion, an oil palm model called Sawit was
developed and has been parameterized to simulate the growth and yield of oil
palms under the influence of weather conditions, planting densities and soil
textures. Improving the representation of oil palm microclimate and plant water
uptake under dense canopies, incorporating fruiting activity, and refining the
trunk's dry matter partitioning mechanism could enhance Sawit's accuracy. |
|---|