Climate change had nowadays constituted a serious global environmental threat.
Scientists predict that the enhanced greenhouse effect could intensify climate
variability. This phenomenon can cause severe weather, condition which might
result in a natural disaster. The damage ensuing from climate change could result
in invaluable losses to a communitys resources such as human, material,
economic as well as environmental adversities (UNEP, 2011).
Agricultural activities have been most vulnerable due to climate change (IPCC,
2001). Consequently, changes in temperature and rainfall for instance, could
significantly affect the yield of agricultural farms (Murad
et al., 2010). Sequel to the above, the decline in farm yield would
impact unfavorably on earnings from agricultural output.
It is contended in existing studies that a rise in the incidences of heat stress,
droughts and floods had far-reaching effect on agricultural productivity. The
results of mean climate change, has negative economic impacts on the agricultural
sector. Climate change will have agronomic impacts on crop yield which will
consequently have economic effects on agricultural prices, production, demand,
trade, regional comparative advantage and producer and consumer welfare (Li
et al., 2011). For instance, a study by Vaghefi
et al. (2011) has reported that a rise in temperature by 2°C
on the Malaysian rice production consequences a reduction of 0.36 t ha-1
which approximately causes economic loss of RM 162.531 million per year under
such climate scenario.
The inherent risk in supply instability for export earnings on plantation crops
as well as the threat to national food security constitutes the core of climate
change concern in Malaysia. Any unfavorable climate will negatively affect agricultural
growth (Murad et al., 2010). Therefore, climate
change phenomenon is an important issue that should be taken into account in
maintaining the sustainability and productivity of agricultural crops. There
are various measures for crop cultivation which could be employed to adapt to
the current climate change event in order to minimize crop damage in the event
of unexpected bad weather (Adger et al., 2007).
In order to identify how climate change could negatively impact the Malaysian
socio-economy, it is necessary to understand the nature of climate variability.
The description of the changing pattern of the climate could be understood by
analyzing the pattern of daily temperature and rainfall for certain a period
of years (Al-Amin et al., 2011).
The objective of this study is threefold. Firstly, to access the historical impact of rainfall and temperature variability on palm oil production. Secondly, to estimate the marginal impact of change in the context of rainfall and temperature on palm oil production. Thirdly, to project the potential long-term future impact of climate variability on palm oil net revenue.
Oil palm is best suited to the humid climate in Malaysia where the rain occurs
at night and the days are bright and sunny. For optimum yield, the minimum rainfall
required is around 1,500 mm year-1 with an absence of dry season
and an evenly distributed sunshine exceeding 2000 h year-1 (Basiron,
2007). A mean maximum temperature range between 29 to 33°C and minimum
temperature between 22 to 24°C favor the highest oil palm bunch production
(Corley and Tinker, 2003).
A high mean annual temperature that falls within the specified weather conditions
required for the thriving of oil palm is favorable for higher fruit production.
However if such high temperatures will lead to drought conditions, it is estimated
that 208,000 ha of land or 12% of the present area would be considered as either
marginal or unsuitable for oil palm cultivation. Lack of water can cause decreased
production of female and male flowers. Besides, drought can also cause flower
drop and reduce crop production (Corley and Tinker, 2003).
Goh et al. (2002) reported that heavy intensity
of rainfall can could result in decreased Fresh Fruit Bunch (FFB) yields, which
will affect severity and frequency of moisture stress experienced by the palms.
The harvest activities will be delayed and consequently the fruit of oil palm
will rot if left for a long time.
According to MMD (2009), the climate change experienced
in Malaysia is driven by two main phenomena which are El-Nino and La-Nina. El-Nino
denotes hot and dry weather while La-Nina is used to describe wet weather characterized
by excessive rainfall. Since the last 30 years, the two strongest El-Nino that
have occurred were experienced in the year 1982-1983 and 1997-1998. These scenarios
could be observed in Fig. 1. This severe climate has effect
on the yield of crop oil palm production. The impact of the El-Nino-driven drier
weather can be noticed now with the development of lower oil-yielding male flowers
and can extend until next year.
Recently, the Economics and Development Division of MPOB
(2010) reports that palm oil production declined because the direct and
indirect uncertainties of climate change. The CPO drops by 3.3% to 16.99 MT
from 17.MT in 2009. Peninsular Malaysia and Sabah both registered declines of
6.1 and 2.5% from 9.5 and 5.3 MT, respectively.
|| Annually CPO production yield (total) by region (Tonne ha-1)
effected by El-Nino episodes
However, Sarawaks CPO production has increased by 9.3% to 2.2 MT. MPOB
reported that average FFB yield fell 6.1% while OER lowered by 0.2% in 2009.
Peninsular Malaysia was the worst affected with average FFB yields declining
by 7.5%, followed by Sabah 4.7% and Sarawak 2.6% (MPOB, 2010).
So, the objective of the study is to estimate the potential impact of climate
change on palm oil production in Malaysia.
MATERIALS AND METHODS
Model specification: In order to study and determine the sensitivity
of oil palm production to CC, the Ricardian model which is currently employed
under this study had been proven to be useful by a number of researchers (Mendelsohn
et al., 2001; Mendelsohn, 2007; Gbetibouo
and Hassan, 2004; Seo et al., 2005; Adger
et al., 2007). The Ricardian model estimates total net revenue that
accrues to farms per hectares of cultivated land depending on climate and other
explanatory variables such as price, quantity of goods and temperature (Seo
et al., 2005). This modeling approach to analyze the impact of climate
on crop productivity is based on a set of CC scenarios to measure the impact
of CC on the net revenue that accrue to farmers. Besides, it is employed to
study the impact of climate change on agriculture (Adger
et al., 2007). Ricardian method assesses the performance of farm
by quantifying impacts on agricultural productivity across the landscape revealing
the effects of variation between different climate zones (Hassan,
2010). Below is the general formulation to calculate the net revenue:
where, NR is net revenue per hectare, Pi is the market price of crop i, Qi
is output, X is a vector of purchased input, Z is the economic variables and
Px is a vector of input prices. Y is a vector of climate variables. The study
defines net revenue as the gross revenue less per hectare production cost. The
dependent variable will be regressed on climate and other control variables.
Following previous specification by Mendelsohn (2007),
the standard Ricardian model relies on a quadratic formulation of climate. This
study extends the specification by adding area; which refers to
the area of cultivated land committed to the planting of oil palm. Thus we model
both the basic model and an extended model as specified below:
||Model 1(without control variables):
||Model 2 (with control variables):
In the above model specifications (Model 1 and 2) for both Eq. 2 and 3, where rain is denoted as rainfall, temp is abbreviation for temperature and ε refers to the error term. Both linear and quadratic terms captured for rainfall and temperature are introduced. The introduction of quadratic terms (rain2 and temp2) stand for the nonlinear values of rainfall and temperature. These reflect the non-linear shape of response function between net revenues and climate. This study also extends the specification on Model 2 in Eq. 3 by adding area; which refers to the area of cultivated land committed to the planting of oil palm.
After estimating the model above, simulations were undertaken using different climate scenarios to determine how palm oil production will be effected under the scenario. The sign of the linear term indicate the unidirectional impact of the independent variable, the quadratic term reflects the non-linear shape of the net revenue of the climate response function. When the quadratic term is positive, the net revenue function is U-shaped and when the quadratic term is negative the function is hill-shaped. The expected marginal impact of a single climate variable on the land value and farm net revenue evaluated at mean is:
where, Following Kurukulasuriya et al. (2006)
the specification in Eq. 4 is used to calculate the marginal
impact on Net Revenue (NR) for each rainfall and temperature climate variables
(fi). Marginal impact calculations determine how palm oil production
will be affected by a 1°C rise in temperature and a 1 mm increase in rainfall.
The advantage of this approach is that the method includes direct effect of
climate on yield on crops and control variables as adaptation response.
Control variable: The control variables were introduced to control for
extraneous factors. It is important to clarify the reason for including the
control variables. The control independent variables represent one of the non-climate
features that influence the net revenue (per hectare). Sanghi
and Mendelsohn (2008) reported that control variable is useful for the purpose
of controlling the influence of missing data (variable). In this study, as the
net revenue per hectare is expected to be influenced by factors other than climatic
variables, we have included one control variable which is oil palm area planted.
Data: Oil palm raw data on net revenue and its determinant were collected
from Malaysian oil palm statistics reported by Economics and Development Division
of the Malaysian Palm Oil Board (MPOB). Three regions were selected for oil
palm cultivation, which are Peninsular, Sabah and Sarawak (Fig.
2). Average annual temperature and rainfall from 1980 to 2010 were obtained
from MMD (2009).
Statistical analysis: The data for this study covers Malaysias
three main regions of the Peninsular, Sabah and Sarawak. Annual time-series
data spanning 31 years (1980-2010) was used to analyze the impact of climate
change in each of the aforementioned regions.
||Oil palm estate distribution in Malaysia, Centre of Remote
Imaging, Sensing and Processing (CRISP), Singapore 2010
Using E-views software, a Ricardian model was estimated based on Ordinary Least
Squares (OLS) regression method for each region. The motivation for employing
the Ricardian method was informed by its popular use to estimate economic impact
of climate change in agricultural sector since the fairly recent past. The regression
results support the presence of climate impact as evidenced by Eq.
2 (Model 1) and Eq. 3 (Model 2). Interpretations of results
have been done through an examination of both the signs and magnitudes of estimated
parameters. Sequel to this, the estimation of the marginal impact value based
on fluctuations in temperature (per°C) and rainfall (per mm) was made possible.
RESULTS AND DISCUSSION
Estimated model: This study had estimated a basic model (Model 1) and a variant model (Model 2) by including the area of cultivated land as a control variable. The essence is to consider beyond climate factors, the effects of related but non-climate factors. The basic model only considers climate factors and their quadratic terms as the most essential independent variables. Meanwhile the extended model (Model 2) examines how the inclusion of the non-climate factor (planted or cultivated area) might impact the dependent variable (net revenue) as well as the statistical significance of the climate variables. The climate variables considered (temperature and rainfall) were each squared, resulting in a quadratic (non-linear) model. This was done to reflect the non-linear shape of response function between net revenues and climate.
Table 1 presents the result from the regression with net
revenue value as dependent variable. Rainfall as well as its squared effect
is 5% level of statistically significant for Peninsula region in both Models
(Model 1 and 2) but statistical evidence does not support the impact of temperature
on net revenue. Both Models reveal that the relationship between rainfall and
net revenue is U-shaped (non-linear). In addition, adjusted R2 value
shows that Model 2 (with a magnitude of 0.743) is better than Model 1 (which
has a smaller value of 0.729). According to MMD (2009),
heavy rainfall is experienced during wet season, from about November to March.
|| Determinants of net revenue per hectare in Peninsular
|*,**,***Significant at 10, 5 and 10%
|| Determinants of net revenue per hectare (Sabah)
|*,**,***Significant at 10, 5 and 10%
This is induced by the east winds from the South China Sea to the east coast
of the Peninsular. Meanwhile, the south-west wind blows from the west Sumatra
Peninsular. Records show that heavy rains that cause floods occur within months
of each year in the monsoon season. This caused the flooding to occur in some
places which take few days to recede. During flooding periods, fertilization
as well as harvesting activities is hindered reducing outputs. This result,
is coherent with a previous finding by Goh et al.
(2002), who studied oil palm grown in coastal alluvial soils in peninsular
Malaysia. They suggested that the warmer climate itself does have significant
influence on production.
Meanwhile, from the result in Table 2, both climate indicators
of rainfall and temperature as well as both squared effects in both models,
were found to have 5% level of statistically significant impact on net revenue
in Sabah region. In addition, R2 shows that much improved Model 2
(0.777) than Model 1 (0.514). Meanwhile, Model 2 shows that the control variable
(planted area) also impacts on net revenue. However, a per cent increase in
cultivated area/ha reduces net revenue by a small amount of 0.1%. This is because
flooding which result from rain variation reduces farm output by delaying fertilization
and harvesting. An often cited example in support of this is the 1996 Greg storm
event. Another example is that of Hilda tropical storm, which brought heavy
rain to Sabah in year 1999 (MMD, 2009). This rain caused
terrific floods and landslides. This havoc was exacerbated by the geographical
positioning of Sabah region which exposed it to winds from the neighboring country,
namely, South China and Philippine. The wind typically brings heavy rains and
storms flooding several days, especially in low areas (MMD,
|| Determinants of net revenue per hectare (Sarawak)
|*,**,***Significant at 10, 5 and 10%
The report also portrays that the El-Nino and La-Nina events have greater influence
on rainfall reduction in East Malaysia (especially in Sabah) compared to peninsular
Malaysia. This evidence while rainfall and temperature are important indicators
which affect palm oil net revenue in Sabah.
In the peculiar case of Sarawak region, statistical evidence supports the impact of temperature on net revenue as both the level and quadratic terms are found to be 1% level of statistically significant at reducing palm oil revenue. The relationship between temperature and revenue is inverted U-shaped. In the case of the control variable, although there is statistical evidence that cultivated area of land increases palm oil revenue, the magnitude is small since one per cent increase in cultivated area/ha will increase revenue only by 0.0053%. Despite this impact of temperature on palm oil revenue, the study finds no relationship between rainfall variations and palm oil revenue. Moreover, R2 seems like unchanged by given Model 2 (0.387) compare to Model 1 (0.380). This is understandable since relatively, Sarawak is geographically sheltered from east winds, especially from Philippine which usually result in storm clouds that consequently cause heavy rain. This explains why heavy rains that cause terrific floods are rarely encountered in this region. In addition, low moisture caused by high temperatures in the dry season in Sarawak has affected the volume of oil palm fruits grown.
Overall, in all regions as shown in Table 1-3, climate variability indicated by either rainfall or temperature or both, are found to impact negatively on palm oil revenue. Model fits are better in Model 2 compared to Model as the values of adjusted R2 improved after adding the control variable and it is found to be statistically significantly in all regions.
Nonetheless, the magnitude of the coefficients representing the impact of cultivated area on palm oil revenue is practically insignificant because the values are very small and hence, negligible.
This suggests that the control variable has relatively smaller impact on oil
palm net revenue when compared to the climate variables. Moreover, the result
is consistent with the report by Mahmood et al. (2007)
that the price factor is a significant variable explaining area changes which
have significant positive coefficient relation of lagged acreage. Future research
could focus on other control variables which might have more spectacular impact
on oil palm production in the face of climate change.
Marginal impact analysis for net revenue model: The marginal impact
analysis was done to assess temperature and rainfall impacts on net revenue
per hectare of oil palm production. Table 4 shows the marginal
effect of climate change on net revenue (RM/hectare). Particularly shown by
Model 1, the marginal effect of rainfall for every 1 mm increase will reduce
net revenue by RM 4.59 and RM 1.60 for Sabah and Sarawak, respectively.
|| Marginal effect of temperature and rainfall on net revenue
|| Average estimated net revenue (RM ha-1) under
PRECIS climate scenario
Meanwhile average peninsular region gains increase in net revenue for every
increase per 1 mm rain by RM 5.24. On the other hand, the marginal effect of
temperature on net revenue presented by Model 1 shows that, for every 1°C
increase, revenue reduces by RM 44.52, RM 45.60 and RM 37.70 for Peninsular,
Sabah and Sarawak, respectively. Based on the total average value from results
obtained in both Model 1 and 2, the highest impact of climate (temperature and
rainfall) variability on palm oil revenue is obtained in Sabah Region. Meanwhile
the least effect is observed in Sarawak region.
Future climate impacts: This study had projected the impact of climate change on oil palm production net revenue (RM/ha) by applying PRECIS Regional Climate Modeling System on oil palm production. The study simulated the impact of future climate change scenarios by using estimated coefficients for net revenue function. We estimate the consequences of climate change scenario for the years; 2029, 2059 and 2099 shown by result in Table 5. By projection using regional climate modeling system, PRECIS, palm oil is predicted to lose an average (RM/hectare) amount of RM 341.29, RM 127.43 and RM 51.80 for Peninsula, Sabah and Sarawak, respectively in year 2029. This means that Peninsular shows highest affect compared to Sabah which is the lowest affect of future predicted climate change. The predicted loss in net revenue rises over the years if no adaptation and mitigation strategies are taken. The essence of this excise is to examine the role climate change might play in the future.
The study attempted to investigate the economic impacts of climate change on
oil palm production in three regions, namely Peninsular, Sabah and Sarawak.
The study applies modified Ricardian method to estimate the effect of climate
change on oil palm production net revenue in Malaysia. Based on the objectives
of this study, the relevant conclusion is summarized as follow; firstly, increase
in rainfall and high temperature will negatively impact on palm oil production.
As such net revenue from oil palm production will reduce annually. Besides,
in the impact of climate variability on output will be more apparent in the
future if adaptation measures are not taken. This study thus, suggests that
adaptation and mitigation strategies are taken in order to minimize the adverse
effects of climate change on oil palm production.
The author would like to thank Universiti Putra Malaysia for financial support provided for this work under Graduate Research Fund with grant numbers: UPM/SPS/GS26921.