Beruflich Dokumente
Kultur Dokumente
discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/310586136
CITATIONS READS
2 305
12 authors, including:
Some of the authors of this publication are also working on these related projects:
Trade-offs and synergies in climate change adaptation and mitigation in coffee and cocoa systems
View project
Collaborative Research Centre 990: Ecological and Socioeconomic Functions of Tropical Lowland
Rainforest Transformation Systems (Sumatra, Indonesia) View project
All content following this page was uploaded by Christopher Richard Donough on 13 February 2017.
Agricultural Systems
a r t i c l e i n f o a b s t r a c t
Article history: Narrowing the gap between actual and attainable yields in existing oil palm plantations is perceived as a key to
Received 2 March 2016 fulfill the growing global demand for vegetable oil. To assess the scope for intensification we need robust esti-
Received in revised form 11 November 2016 mates of attainable yields, which has been so far rarely done for perennial crops. For this purpose, we evaluated
Accepted 14 November 2016
the complexities associated with estimating yield gaps (YGs) in oil palm (i.e. carry-over effect and aging), and
Available online xxxx
adapted the existing framework for YG studies in annual crops. Based on this framework, we analyzed YGs for
Keywords:
four sites within oil palm plantations located in Sabah (Malaysia), Central Kalimantan and North Sumatra (Indo-
Yield gap assessment nesia) using a unique commercial yield data set covering an area of 38,300 ha.
Oil palm We assessed for each site at plantation scale water-limited potential yield using the PALMSIM simulation model,
Carry-over effects attainable yield determined by best performing blocks within the plantation as defined by 90th percentile of ob-
Attainable yield served yields and actual yields (blocks representing the median yields). The water-limited potential yield did not
differ very much; 35–39 t fresh fruit bunch (FFB) during the plateau phase, the most productive phase in the life
time of a palm. This reflected the favorable environmental conditions found in many parts of Sumatra and Borneo
for oil palm. Attainable yields were in the range of 26–31 t FFB/ha. The exploitable YG between attainable and ac-
tual yield ranged for the four sites from 5 to 7 t FFB/ha/year. For one site (Central Kalimantan), we assessed yield
variability due to varying soil conditions at the block scale according to its dominant soil type. This suggested that
they were indeed exploitable by management. If the plantation could close the gap between attainable and actual
yield this could give about 21,000 t/yr higher FFB. This indicated the large scope for intensification oil palm offers
in many parts of insular Southeast Asia.
© 2016 Elsevier Ltd. All rights reserved.
1. Introduction concerns about the conversion of tropical forests into plantations of oil
palm. These include increased CO2 emissions from deforestation and
Global production of vegetable oil (palm, soy, canola and sunflower) degradation of peat soils, and loss of biodiversity (Carlson et al., 2012;
doubled from 73.9 Mt to 141.0 Mt 2000–2013. Production of palm oil in- Koh and Wilcove, 2008). Environmentalists are often at odds with fron-
creased from 25.0 to 61.1 Mt (2.5 times) over the same period tier developers, particularly in Southeast Asia (Sayer et al., 2012). There
(FAOSTAT, 2015). Oil palm (Elaeis guineensis) can produce up to is a global movement to reduce deforestation and restrict new oil palm
10 t/ha/yr crude palm oil (CPO) in favorable sites (Corley and Tinker, plantations to low-carbon, degraded land, including land that was
2016). Its genetic yield potential is 11–18 t CPO/ha/yr (Barcelos et al., cleared from forest in the past (Gingold et al., 2012).
2015). As global demand for vegetable oils increased, the oil palm indus- An alternative to expanding the area of oil palm is to increase the
try responded by expanding the planted area. The area planted to oil productivity of existing plantations (Garnett et al., 2013). But first we
palm in Indonesia, the world's major producer of palm oil, doubled be- need to know how actual yields compare with those that can be obtain-
tween 2003 and 2012 (FAOSTAT, 2015). This has led to environmental ed with good management, which is called the yield gap (YG) (Connor
et al., 2011, p.11). Most analyses of YG have been with annual crops. In
⁎ Corresponding author. their analysis of global crop yields and global food security, Fischer et al.
E-mail address: mhoffma@gwdg.de (M.P. Hoffmann). (2014) focus on YGs in food crops, including oil palm. They cite 1011
http://dx.doi.org/10.1016/j.agsy.2016.11.005
0308-521X/© 2016 Elsevier Ltd. All rights reserved.
M.P. Hoffmann et al. / Agricultural Systems 151 (2017) 12–19 13
references, of which 554 are for annual crops, 448 we could not identify initiation to harvest is somewhat over three years (Breure, 2003). The
from the reference title, but only 9 were for perennial crops. A further sex of the inflorescence differentiates about two years before harvest,
literature search found little work to identify YGs in any perennial when the potential number of flowers is fixed, although inflorescences
crop, far less a perennial crop as important as oil palm. may abort about 10 months before harvest. Pollination occurs about
The aim of this paper was first to outline the complexities associated six months before harvest and the potential number of fruits is deter-
with estimating YGs in oil palm, and to adapt the existing frameworks mined, although some flowers may abort later while the remaining
for YG assessments to oil palm. Based on the revised framework, we fruits develop.
identified YGs for four sites in plantations in Indonesia and Malaysia. During the period from initiation to harvest, variations in rainfall in-
We then explored yield variability due to palm age and soil type. We fluence the development of the fruit bunches. Stress prior to pollination
used a data set of yield records from 1198 management blocks covering reduces the sink size, while stress after pollination reduces growth of
an area of 38,300 ha for several years. These data provided insights into the bunch (Legros et al., 2009a, 2009b). Estimates of PYw must there-
current productivity of commercial plantations and the scope for im- fore consider the rainfall quantum and its distribution over the three
provement in Southeast Asia. years prior to harvest.
Cock et al. (2016) showed that extremes of rainfall during the last
2. Yield gap analysis in oil palm two years before maturity reduces yield, but we know little about the
long-term carry-over effects of water stress on oil palm. A common ob-
The first step to analyze YGs was to determine the potential yield servation in field trials is that a year of high yields is followed by one or
(PY), which is the “yield to be expected with the best-adapted variety more years of low yield (Breure and Corley, 1992; Corley and Tinker,
(usually the most recent release), with the best management of agro- 2016). Commercial operators confirm this, attributing the yield decline
nomic and other inputs, and in the absence of manageable abiotic and in 2010 in one case to the high levels of production in 2008 and 2009
biotic stresses” (Fischer et al., 2014, p. 30). Oil palm is rarely irrigated (United Plantations, 2010). This can be because the number of bunches
in the tropics where most is currently grown, although we are aware and the number of fruit per bunch is set by the weather and yield in the
that irrigated oil palm is an option in native savannahs with a dry sea- two years or so between flower initiation and pollination. Oil palm can
son. As we focus attention on intensification in established plantations therefore yield well in years with dry spells by mobilizing carbohydrate
we step down to water-limited potential yield (PYW), which is “the reserves from the trunk during fruit development. Trunk reserves can
yield obtained with no other manageable limitation to the crop apart contribute up to 5 t/ha FFB (Henson et al., 1999; Henson and Dolmat,
from the water supply” (Fischer et al., 2014, p.32). This puts the YG in 2004). There may also be long-term effects from nursery management,
the context of the climate and soil. Attainable yield (AY) is “the yield but there is little evidence to support this hypothesis.
attained by a farmer from average natural resources when economically
optimal practices and levels of inputs have been adopted while facing 2.2. Yield taking
the vagaries of weather” (Fischer et al., 2014, p. 32). In contrast, farm
yield (FY) is “the field, district, regional or national average yield given The yield taken (fruit harvested) in oil palm is often a lot less than
in kilograms or metric tonnes per hectare (kg/ha and/or t/ha).” AY the yield made (fruit produced) due to practical difficulties at harvest-
with no subscript implies well-watered or irrigated crops, while AYW ing (Cock et al., 2014). Thus, the yield data from commercial plantations
is for rainfed crops. “FY … for all countries are collated annually by … often underestimate the actual yield produced. Moreover, the reliability
[FAO] and are disseminated through the publicly accessible database of data of taken yield at the block level depends on how it was done.
FAOSTAT” (Fischer et al., 2014, p. 30). The YG is the difference between Plantations usually record the number of bunches harvested from
FY and AYW for the site and crop cycle under consideration. It is each block, but they are not weighed. Bunches from several adjacent
expressed as a percentage of FY since this “is the observed world … pro- blocks are collected for transport to the mill where the load is weighed
duction and likely increases are directly linked to FY (not PY)” (Fischer and the average bunch weight of the whole load calculated. This average
et al., 2014, p. 33). weight is applied to the number of bunches collected from the relevant
While the growth cycle of tropical annual crops rarely exceeds six blocks to determine the total yield of each block. Because each load may
months, commercial oil palm has a sequence of production cycles over come from several blocks whose mean bunch weight may differ, the cal-
at least 25 years. Soil, terrain and plant genetics are fixed for this time culated block yield may not be accurate. Moreover, the interval between
and can be analyzed in the same manner as for annual crops. In annual harvest and weighing can very a lot, causing further errors. Neverthe-
crops, crop performance is affected by variation in weather patterns less, errors in the yield of individual blocks are irrelevant in calculating
over the growth cycle of less than six months. In contrast, oil palm fruits taken yields of whole plantations or estates. They are sufficient to iden-
take more than three years from flower initiation to harvest maturity. tify the outcome of what good management can achieve in a given
Thus, for example, a dry period at any time over these three years, in- region.
cluding the harvest year, may drastically reduce yield. Similarly, any re-
duced applications of fertilizer over the several years before harvest will 2.3. Age effects
decrease yield. These long-term carry-over effects related to the complex
physiology of oil palm must be taken into account when assessing YGs. Palms yield less as they age. They start to produce bunches three
In addition to these carry-over effects from year to year, harvested years after planting with the yield increasing rapidly to a plateau
yields commonly decline as the palms age. Moreover, the records of har- phase that starts when the palms are 6 to 7 years old lasting until they
vest yield (yield taken) do not always reflect independent estimates of are 10 to 12 years old. Increasing yield in the early phase is associated
the fruit available for harvest (yield made) (see Section 2.2 below). with increasing leaf area index and canopy closure, which leads to
These factors, combined with differing scales of assessment, from small- greater interception of incident radiation and hence yield. After canopy
holder plots to plantation blocks (20–100 ha) to estates (collections of closure, yields reach a plateau for a number of years, after which pro-
20–50 blocks) or plantations (collections of estates) make addressing ductivity declines. Part of the yield decline is attributed to loss of
YGs in oil palm a challenge. stand, mainly due to diseases, as well as lower yield per palm (Goh et
al., 1994). Declining yield may be because old, tall palms are more diffi-
2.1. Carry-over effects cult to harvest and prune (Goh et al., 1994) or because maintenance res-
piration of the larger trunks of older palms is greater (Henson, 2004).
Fresh fruit bunch (FFB) yield of oil palm is determined by the num- Respiration in other crops is closely related to total photosynthesis
ber of bunches per ha and their mean weight. The period from flower and growth (Cheng et al., 2009; Frantz et al., 2004; Thornley, 2011). If
14 M.P. Hoffmann et al. / Agricultural Systems 151 (2017) 12–19
this is so, increased maintenance respiration is unlikely to be a major of assessment require more comprehensive data, which makes site-spe-
cause of yield decline as palms age. Whatever the reason for the decline cific assessment and hence management more complex (Fig. 1).
in taken yield, it is essential to take account of stand age in assessing
PYw of oil palm (von Uexküll et al., 2003).
2.5. Yield gap framework
2.4. Scale of assessment The yield-gap framework was developed for annual crops (van
Ittersum and Rabbinge, 1997; van Ittersum et al., 2013; Fischer et al.,
The scale of the YG analysis depends on the purpose of the study. 2014). It does not represent the carry-over and age effects required for
Policy makers might be interested in the scope for intensification at na- perennials like oil palm. We suggest a complementary framework
tional or provincial scale. Companies seeking investment want to know based on that for annuals, but linked to a dynamic component that
about the gap at plantation sites, while plantation managers might want takes account of plant age (Fig. 2). We retain the concepts of PY, PYw
to assess the scope for each block. In commercial oil palm plantations, and FY yield as defined for annual crops, but incorporate the concept
the block, usually 25 ha or more, is the smallest unit that receives uni- of variation of yield with plant age.
form management. This does not take account of palm-to-palm varia- Commercial oil palm is grown in equatorial climates with annual
tion, which can be substantial in non-clonal blocks (von Uexküll et al., precipitation N1600 mm and rarely irrigated (Carr, 2011). We therefore
2003). This variation is due to both the genetic variation between indi- ignore the difference between PY and PYW in Indonesia and Malaysia.
vidual palms and soil and topography, which can vary widely within Estimates of PY and PYW are mostly from simulation modelling (van
one block. PYw can therefore differ substantially within one block due Ittersum et al., 2013).
to plant and spatial variability. Furthermore, management practices, Fischer et al. (2014) focused on national and regional differences and
such as fertilizer applications, may be sub optimal for some palms and their influence on global food security. We use their concepts to help an-
supra optimal for others within a single block. alyze performance of oil palm at the level of the plantation block. We de-
At the national and regional level, yield ceilings are driven by climate fine YG in oil palm (YG) as the difference in yield between an optimally
(Fig. 1). At decreasing scale, palm age, plant genetics and soil properties managed and a sub-optimally managed block (Fig. 3). Better agronomic
become important to explain variation in FY and AY. Management ex- management can close this gap. We divide YG into four sub-gaps, based
plains variability in the actual yield at the plantation and the block on Fairhurst and Griffiths (2014). YGE represents nursery management
level, while genetic differences operate at the level of individual plants. and plantation establishment while YGN represents sub-optimal nutri-
Note, however, that vegetatively-propagated palms are genetically ent management. YGA represents other agronomic management op-
identical. The scale of assessment therefore dictates the methodologies tions such as weeding, pruning, ground cover, mulching and so on.
needed to define and to explore yield levels. It follows that finer scales YGY represents harvest efficiency, the difference between yield made
Fig. 1. Methodology and corresponding information needed to assess yield gaps in oil palm. PALMSIM (Hoffmann et al., 2014) and APSIM (Huth et al., 2014) are crop models. BMP is the
acronym for Best Management Practice (Fairhurst and Griffiths, 2014).
M.P. Hoffmann et al. / Agricultural Systems 151 (2017) 12–19 15
reflects typical current growing regions of oil palm. Table 1 gives an over-
view about site conditions and yield recordings. For rainfall, we used the
NASA data (NASA, 2016) instead of the actual measured values because
of missing data and possible non-uniformity in data collection method.
Plantation companies classify the soil into soil management groups
(SMGs, Table 2). Soil surveyors usually define the SMGs grouping to-
gether soil types based on similarity of management requirements,
and the dominant SMG for each block is used to guide soil and water
management practices in plantations.
We selected four commercial oil palm plantations. Two were located For site 4, we analyzed the potential to increase production in the pla-
in Central Kalimantan, one in North Sumatra and one in Sabah. This teau phase (7–12 years after planting) from FY to AYw for each SMG. We
100
Yield potential
of progeny
YGE for
given soil
and climate
90
Yield reduced
because of
YGN poor
Relative bunch yield (%)
plantation
establishement
80
Yield reduced
because of
YGA incorrect
nutrient
diagnosis
70
Yield reduced
because of
YGY inadequate
crop agronomic
mangement
60
Yield reduced
because of
inadequate
crop recovery
50
Table 1
Overview about the data set available for this study.
Site Region Number of blocks Period of recording (years) Total area (ha) Rainfalla (mm/year)
multiplied the FY for each SMG with the area covered by the SMG; and as- respectively, FY 24.7 and 25.9 t/ha and YG 24.3 and 12.1%, respectively).
sumed that 70% of the area would be in the plateau phase. We did the Site 3 had lower AYw (25.6 t/ha) and FY (21.0 t/ha) for a YG of 21.9%.
same for AYw and calculated the difference between FY and AYw. Site 1 had the largest YG of 33.3% (AYw 28 and FY 21 t/ha).
4.1. Yield-age profile (YAP) Many plantations have soil maps indicating the dominant soil type
for a block. We used this information for site 4 to analyze YGs according
The plantation data show characteristic YAP patterns of oil palm with to the SMG of each block. In site 4, SMG B and C dominate with about
some important differences between sites (Fig. 4). The juvenile phase 10,100 and 4500 ha, respectively (Fig. 6a). Mean yields between the
went up to ~6 years, where yields steeply increase until they reach the two SMG were not much different over time (Fig. 6b). SMG A and D
plateau phase, the most productive period during the lifetime of the cover about 1045 and 1700 ha respectively and both performed poorly
palm. The rate at which yield climbs to the plateau varies, for example during the juvenile phase, but caught up somewhat during the plateau
between sites 3 and 4, both in Central Kalimantan. The YAP identifies phase. Because the blocks of the plantation were planted in stages
the YG over the production life of the plantation, which is cumulative. over a number of years, there are fewer data for the right-hand ends
There was higher variability in yields up to year nine after planting of the curves for each SMG. In the case of SMG D at yr 10 there were
after which the variability was reduced. We superimposed the simulated 22 blocks with median 21.6 t/ha; at yr 11, 6 blocks, median 25.8 t/ha;
values from PALMSIM. For site 1, 2 and 4 the model underestimated yields but at yr 12 there were only 2 blocks, median 28.5 t/ha. We therefore
in the steep ascent phase. One possible reason could be is that it is a com- disregard the ‘uptick’ for SMG D in yr 12.
mon practice to report ‘yield per harvested hectare’ in the first year, espe- Differentiating the YG for the plateau phase for the four SMG (Fig. 7)
cially in blocks where only a part of the block is harvested first. This will showed that AYw and FY did not differ much between SMG (Fig. 7) in
artificially inflate 1st year production records. However, the age-specific this plantation. AYw ranged from 32 to 29 t FFB/ha and FY from 24 to
trend is the same in observed as well in the simulated yields. 28 t FFB/ha. The YG was about 10% in SMG A, B and C and 19% in SMG D.
In a next step we used these YGs for the plateau phase and calculated
4.2. Yield gaps for plateau phase a scenario where all blocks within the plantation achieved their SMG-
specific AYw. We compared this output with a scenario where all blocks
We estimated PYw, AYw and FY for the four plantation sites. To ac- only achieved their actual FY. We assumed that 70% of the plantation
count for year-to-year -variation and age effect (see Section 2.3) we area would be in the plateau phase. The difference in FFB between all
used average t FFB/ha/year during the plateau phase (7–12 years after blocks at SMG-specific FY and AYw would be 21,000 t/yr for the
planting), instead of maximum reported yield for a single year. Simulat- whole plantation, representing about US$3.15 M (at 25% oil per t FFB
ed PYw during this period did not differ very much ranging from 35 to and US$600/t oil).
39 t/ha (Fig. 5). AYw defined by the best performing blocks reached
88 and 86% of PYw at site 2 and 4. Site 1 and 3 achieved only 78 and 5. Discussion
66% of PYw.
The YG is that between the yields of the best performing blocks (AYw) 5.1. Yield gaps and scope for intensification
and the median yields (FY). It is this gap that is potentially closable by im-
proved management for the same SMG (Figs. 2 and 3). Average AYw and To remove the carry-over effects and the effects of palm age and
FY were already high for sites 2 and 4 (AYw 30.7 and 30.6 t/ha weather (Henson and Harun, 2004), we examined the means of actual
Table 2
Soil management groups and their associated soil types according to FAO classification found in the four study sites.
Fig. 4. Fresh fruit bunch yield (t/ha) plotted against years after planting for the four plantation sites. Boxes are the interquartile range (IQR) and the white line is the median. The whiskers
are 1.5 × IQR. The line is PYw simulated by PALMSIM.
yields for 7–12 years after planting. The analysis showed marked differ- addition, between-year variation was much greater at site 1 compared
ences in YGs between the four sites. The estimated AYw was 26– with site 4 (Fig. 4). Further analysis is needed at this scale to identify
31 t FFB/ha. Site 1 is doing poorly with a YG of over 33% (Fig. 5). Sites the specific reasons. Nevertheless, YG analysis for the plateau growth
2 and 3 are intermediate with YGs of 24 and 22%, and have room to im- phase is useful to assess current performance at the plantation scale
prove. Site 4 is doing quite well with a YG of just 11%, yet the plantation and compare plantation sites with similar climates.
forgoes about US$3M of gross annual income, detailed below. In Our conclusion that we had to include age-specific yield ceilings to
assess YGs agrees with Euler et al. (2016). They found that smallholders
in Jambi (Sumatra, Indonesia) achieved FYs close to AYw in the post-
plateau (declining) phase, but there were large YGs during the plateau
phase. These were caused by lack of labor and fertilizer, which could
be narrowed by increasing inputs.
PYw ranged between 34 and 39 t/ha/yr reflecting that the climate is
favorable across the major production regions for oil palm in insular
Southeast Asia. This compares with West Africa, the center of origin of
oil palm and where it was originally grown in plantations (Rhebergen
et al., 2014, 2016).
Estimates of PYw rely on simulation modelling, which must be used
with care due to uncertainties in model parameterization and in the cli-
mate data. As discussed above, crop modelling for perennials is still in its
infancy in comparison to annual crops. For example, the current version
of PALMSIM ignores effects of temperature and flooding. Furthermore,
the climate data we used are from the NASA database, not from the
sites. Nevertheless, the simulated PYw data are close to other estimates
for the major production regions of Southeast Asia, which supports its
plausibility (Fairhurst and Griffiths, 2014).
We divided the yield data for site 4 according to the dominant SMG
for each block (Fig. 6a). There was little yield variation between the
two dominant soils SMG B and SMG C (Fig. 6b). In contrast, SMG A and
SMG D yielded less in the juvenile phase although SMG A did improve
during the plateau phase. SMG D was always inferior. The similarity of
Fig. 5. Assessed production level for each plantation: water-limited potential yield (PYw) SMG B and SMG C indicates that much of the variability between blocks
for each plantation simulated by PALMSIM and attainable yield (AYw; 90th percentile of
recorded annual block yields), 75th percentile, farmer yield (FY, median) and 25th
is due to management rather than soil constraints. We calculate that clos-
percentile. Yields as fresh fruit bunches (FFB) (t/ha/year) were averaged for the plateau ing the YG could increase FFB production for the plantation by 21,000 t/yr
phase (7–12 years after planting). with a gross value of US$3.15 M (at 25% oil in FFB and oil at US$600/t).
18 M.P. Hoffmann et al. / Agricultural Systems 151 (2017) 12–19
Fig. 6. (A) The location of the different soil management group (SMG) within the plantation. For explanation of SMG we refer to Table 2. (B) Fresh fruit bunch (FFB) yield (t/ha) averaged
across blocks for each of the four SMG found in site 4 plotted against years after planting.
Other reports confirm that at both the plantation level and the block responses to the rates of fertilizer applied. They found substantial differ-
level, management is the dominant reason for YGs (Cock et al., 2016, ences in response to fertilizer.
Donough et al., 2010). Euler et al. (2016) report similar results for oil
palm grown by smallholders. Harvesting more of the fruit that the 5.2. Methodological issues to improve yield gap assessment/data availability
palms produce appears to increase yield at the plantation level. Called
improved yield taking, palms are harvested at intervals of 10 days or Managers need robust estimates of AYw as proxies to assess YGs that
less, which reduces the substantial losses due to late harvesting they can exploit. The simple framework we introduced above identifies
(Donough et al., 2010). Like most activities in oil palm plantations, har- YGs at the plantation and the SMG level, which managers could exploit.
vesting is by hand, so that the cost and availability of labor are major The scale of the YG assessment must therefore be chosen that will allow
constraints. More mechanization, coupled with breeding dwarf varieties FY to move closer to AYw as affected by the soils x genetics interaction,
to ease harvest and frond management could help overcome these chal- taking into account carry-over effects. Analysis above the block scale,
lenges (Lee et al., 2015). In addition, better infrastructure (roads and however, hides a lot of this variability, which can lead to inefficient re-
bridges) could also make management easier. source use. Field trials and detailed crop models are necessary tools to
Plantations, currently, do not measure, much less monitor, yield re- assess AYw. In our simulation analysis, we took account of only radia-
sponse to applied fertilizers. The Plantation Intelligence analysis platform tion, rainfall and plant age, which misses the carry-over effects of last
(Cook et al., 2014) uses data of commercial operations to estimate yield year's yield. The APSIM palm model (Huth et al., 2014), which needs
more detailed input data than PALMSIM, may be useful in this regard,
but needs further testing to assess its relevance for Southeast Asia.
In the future, the oil palm industry will face socio-economic pressure
due to limited land, volatility in the price of CPO, availability and cost of
labor, and the cost of fertilizer. It will therefore be increasingly chal-
lenged to increase resource-use efficiency, a key component of which
will be matching inputs to increase FY closer to AYw.
Establishment of unfertilized plots (or rows or sub-blocks) in repre-
sentative soil types in a plantation would permit better estimates of the
yield response to applied fertilizers at commercial scale, complementing
the Plantation Intelligence approach (Cook et al., 2014). This defines the
size of YGN (Fig. 3) due to sub-optimal nutrient management. Availabil-
ity of unfertilized yield data also permits estimates of fertilizer-use effi-
ciency beyond the most basic level of yield per unit nutrient applied. It
further allows monitoring of nutrient performance over time as part of
wider sustainability management of the production system.
Better weather and soil data would contribute to a more detailed
yield-gap assessment. A more detailed assessment of the NASA or the
tropical rainfall measuring mission data with oil palm would therefore
be worthwhile. There is also a need to explore the effect of climate ex-
tremes on oil palm caused by the El Niño-Southern Oscillation phenom-
enon and in the future by climate change.
There are few useful data from field trials with oil palm in the liter-
ature, although plantation companies do conduct field trials to aid the
Fig. 7. Farmer yield (FY) as defined by the median yielding blocks and the attainable yield
(AYw) as determined by the 90th percentile for each soil management group (SMG)
development of better management. An agreement to make their data
within site 4. Both yields are given as fresh fruit bunch (FFB) (t/ha). Only blocks in the more widely available would help to quantify site-specific production
range of 7–12 years after planting were considered to reflect the plateau phase. ceilings. To sum up, more work is needed to quantify better site-specific
M.P. Hoffmann et al. / Agricultural Systems 151 (2017) 12–19 19
AYw. Nevertheless, field data showed that there is a substantial exploit- Frantz, J.M., Cometti, N., Bugbee, B., 2004. Night temperature has a minimal effect on res-
piration and growth in rapidly growing plants. Ann. Bot. 94:155–166. http://dx.doi.
able YG for sustainable intensification in oil palm. org/10.1093/aob/mch122.
Garnett, T., Appleby, M.C., Balmford, A., Bateman, I.J., Benton, T.G., Bloomer, P.,
6. Conclusions Burlingame, B., Dawkins, M., Dolan, L., Fraser, D., Herrero, M., Hoffmann, I., Smith,
P., Thornton, P.K., Toulmin, C., Vermeulen, S.J., Godfray, H.C.J., 2013. Sustainable Inten-
sification in Agriculture: Premises and Policies. Sci. Mag. 341:33–34. http://dx.doi.
We presented a framework to quantify the exploitable YG in oil org/10.1126/science.1234485.
commercial oil palm plantation taking in a simple approach to account Gingold, B., Rosenbarger, A., Muliastra, Y.I.K.D., Stolle, F., Sudana, I.M., Manessa, M.D.M.,
Murdimanto, A., Tiangga, S.B., Madusari, C.C., Douard, P., 2012. How to Identify De-
for palm age and carry-over effects. In a second step we applied the
graded Land for Sustainable Palm Oil in Indonesia. Work. Pap. World Resour. Inst.
framework to four plantations located in major production regions of Sekala, Washingt. D.C. (Available online http//wri.org/Publ. Indones. WRI/SE).
Southeast Asia. Moving from plantation to block level illustrated the Goh, K.J., Chew, P.S., Teo, C.B., 1994. Maximising and maintaining oil palm yields on com-
mercial scale in Malaysia. International Planters Conference, pp. 121–141.
large scope for intensification by improved management. A simple cal-
Henson, I., 2004. Estimating maintenance respiration of oil palm. Oil Palm Bull. 48, 1–10.
culation from one plantation showed that closing the YG would result Henson, I.E., Dolmat, M.T., 2004. Seasonal variation in yield and developmental processes
in a higher FFB production of 21,000 t/yr with a gross value of more in an oil palm density trial on a peat soil: 1. Yield and bunch number components.
than US$3M. J. Oil Palm Res. 16, 88–105.
Henson, I.E., Harun, M.H., 2004. Seasonal variation in oil palm fruit bunch production: its
origin and extent. Plant. Kuala Lumpur 80, 201–212.
Acknowledgements Henson, I., Chang, K.C., Siti Nor Aishah, M., Chai, S.H., HAsnuddin, Y., Zakaria, A., 1999. The
oil palm trunk as a carbohydrate reserve. J. Oil Palm Res. 11, 98–113.
Hoffmann, M.P., Castaneda, V.A., van Wijk, M., Giller, K.E., Oberthür, T., Donough, C.,
International Plant Nutrition Institute (IPNI) member company Whitbread, A.M., 2014. Simulating potential growth and yield of oil palm (Elaeis
Canpotex Ltd. supported the research (project number SEA-06). We guineensis) with PALMSIM: model description, evaluation and application. Agric.
thank Dr. Hereward Corley for his comments on an earlier version of Syst. 131:1–10. http://dx.doi.org/10.1016/j.agsy.2014.07.006.
Hoffmann, M.P., Donough, C., Oberthür, T., Castaneda Vera, A., van Wijk, M., Lim, C.H.,
the manuscript. Asmono, D., Samosir, Y., Lubis, A.P., Moses, D.S., Whitbread, A.M., 2015. Benchmarking
yield for sustainable intensification of oil palm production in Indonesia using
References PALMSIM. Plant. 91, 81–96.
Huth, N.I., Banabas, M., Nelson, P.N., Webb, M., 2014. Development of an oil palm
Barcelos, E., Rios, S.d.A., Cunha, R.N.V., Lopes, R., Motoike, S.Y., Babiychuk, E., Skirycz, A., cropping systems model: lessons learned and future directions. Environ. Model.
Kushnir, S., 2015. Oil palm natural diversity and the potential for yield improvement. Softw. 62:411–419. http://dx.doi.org/10.1016/j.envsoft.2014.06.021.
Front. Plant Sci. 6. http://dx.doi.org/10.3389/fpls.2015.00190. Koh, L.P., Wilcove, D.S., 2008. Is oil palm agriculture really destroying tropical biodiversi-
Breure, C.J., 2003. The search for yield in oil palm: basic principles. In: Fairhurst, T, ty? Conserv. Lett. 1:60–64. http://dx.doi.org/10.1111/j.1755-263X.2008.00011.x.
Härdter, R (Eds.), Oil palm: Management for Large and Sustainable Yields 59–98. Laborte, A.G., de Bie, K.C.A.J.M., Smaling, E.M.A., Moya, P.F., Boling, A.A., Van Ittersum, M.K.,
Breure, C.J., Corley, R.H.V., 1992. Fruiting activity, growth and yield of oil palm. II. Obser- 2012. Rice yields and yield gaps in Southeast Asia: past trends and future outlook.
vations in untreated populations. Exp. Agric. 28:111–121. http://dx.doi.org/10.1017/ Eur. J. Agron. 36:9–20. http://dx.doi.org/10.1016/j.eja.2011.08.005.
S001447970002305X. Lee, M., Xia, J.H., Zou, Z., Ye, J., Rahmadsyah, A.Y., Jin, J., Lieando, J.V., Purnamasari, M.I., Lim,
Byerlee, D., Heisey, P., Pingali, P., 2000. Realizing yield gains for food staples in developing C.H., Suwanto, A., Wong, L., Chua, N.-H., Yue, G.H., 2015. A consensus linkage map of
countries: prospects and challenges. Food Needs of the Developing World in the Early oil palm and a major QTL for stem height. Nat. Sci. Reports 5:8232. http://dx.doi.org/
Twenty First Century. Pontifical Academy of Sciences, Vatican City, pp. 207–250. 10.1038/srep08232.
Carlson, K.M., Curran, L.M., Asner, G.P., Pittman, A.M., Trigg, S.N., Marion Adeney, J., 2012. Legros, S., Mialet-Serra, I., Caliman, J.-P., Siregar, F.A., Clément-Vidal, A., Fabre, D.,
Carbon emissions from forest conversion by Kalimantan oil palm plantations. Nat. Dingkuhn, M., 2009a. Phenology, growth and physiological adjustments of oil palm
Clim. Chang. 3:283–287. http://dx.doi.org/10.1038/nclimate1702. (Elaeis guineensis) to sink limitation induced by fruit pruning. Ann. Bot. 104:
Carr, M.K.V., 2011. The water relations and irrigation requirements of oil palm (Elaeis 1183–1194. http://dx.doi.org/10.1093/aob/mcp216.
guineensis): a review. Exp. Agric. 47:629–652. http://dx.doi.org/10.1017/ Legros, S., Mialet-Serra, I., Clément-Vidal, A., Caliman, J.-P., Siregar, F.A., Fabre, D.,
S0014479711000494. Dingkuhn, M., 2009b. Role of transitory carbon reserves during adjustment to climate
Cheng, W., Sakai, H., Yagi, K., Hasegawa, T., 2009. Interactions of elevated [CO2] and night variability and source-sink imbalances in oil palm (Elaeis guineensis). Tree Physiol.
temperature on rice growth and yield. Agric. For. Meteorol. 149:51–58. http://dx.doi. 29:1199–1211. http://dx.doi.org/10.1093/treephys/tpp057.
org/10.1016/j.agrformet.2008.07.006. NASA, 2016. http://power.larc.nasa.gov/cgi-bin/cgiwrap/solar/agro.cgi.
Cock, J., Donough, C.R., Oberthür, T., Indrasuara, K., Rahmadsyah, Gatot, A.R., Dolong, T., Rhebergen, T., Hoffmann, M.P., Zingore, S., Oberthür, T., Acheampong, K., Dwumfour, G.,
2014. Increasing Palm Oil Yields By Measuring Oil Recovery Efficiency From the Fields Zutah, V., Adu-Frimpong, C., Ohipeni, F., Fairhurst, T., 2014. The Effects of Climate,
To the Mills. Int. Oil Palm Conf. (IOPC), Bali. Soil and Oil Palm Management Practices on Yield in Ghana. Int. Oil Palm Conf. 17–
Cock, J., Kam, S.P., Cook, S., Donough, C., Lim, Y.L., Jines-Leon, A., Lim, C.H., Primananda, S., 19 June.
Yen, B.T., Mohanaraj, S.N., Samosir, Y.M.S., Oberthür, T., 2016. Learning from commercial Rhebergen, T., Fairhurst, T., Zingore, S., Fisher, M., Oberthür, T., Whitbread, A., 2016. Cli-
crop performance: oil palm yield response to management under well-defined growing mate, soil and land-use based land suitability evaluation for oil palm production in
conditions. Agric. Syst. 149:99–111. http://dx.doi.org/10.1016/j.agsy.2016.09.002. Ghana. Eur. J. Agron. 81:1–14. http://dx.doi.org/10.1016/j.eja.2016.08.004.
Connor, D.J., Loomis, R.S., Cassman, K.G., 2011. Crop Ecology: Productivity and Manage- Sayer, J., Ghazoul, J., Nelson, P., Klintuni Boedhihartono, A., 2012. Oil palm expansion
ment in Agricultural Systems. second ed. Cambridge University Press, Cambridge, UK. transforms tropical landscapes and livelihoods. Glob. Food Sec. 1:114–119. http://
Cook, S., Lim, C.H., Mohanaraj, S.N., Samosir, Y.M.S., Christopher, R., Lim, Y.L., Cock, J., Kam, dx.doi.org/10.1016/j.gfs.2012.10.003.
S.P., Program, S.A., 2014. Palm oil at the crossroads: the role of plantation intelligence Thornley, J., 2011. Plant growth and respiration re-visited: maintenance respiration de-
to support change, profit and sustainability. Plant. 90, 1–13. fined - it is an emergent property of, not a separate process within, the system -
Corley, R.H.V., Tinker, P.B.H., 2016. The Oil Palm. fifth ed. Wiley-Blackwell. and why the respiration: photosynthesis ratio is conservative. Ann. Bot. 108:
Donough, C., Witt, C., Fairhurst, T., 2010. Yield Intensification in Oil Palm Using BMP As a 1365–1380. http://dx.doi.org/10.1093/aob/mcr238.
Management Tool. International Conference on Oil Palm and the Environment 2010, United Plantations, 2010. Annual Report 2010. http://www.unitedplantations.com/. :
Bali, Indonesia. pp. 1–49.
Euler, M., Hoffmann, M.P., Fathoni, Z., Schwarze, S., 2016. Exploring yield gaps in small- van Ittersum, M.K., Rabbinge, R., 1997. Concepts in production ecology for analysis and
holder oil palm production systems in eastern Sumatra, Indonesia. Agric. Syst. 146: quantification of agricultural input-output combinations. F. Crop. Res. 52:197–208.
111–119. http://dx.doi.org/10.1016/j.agsy.2016.04.007. http://dx.doi.org/10.1016/S0378-4290(97)00037-3.
Fairhurst, T., Griffiths, W., 2014. Oil Palm; Best Management Practices for Yield Intensifica- van Ittersum, M.K., Cassman, K.G., Grassini, P., Wolf, J., Tittonell, P., Hochman, Z., 2013.
tion, First. Edit. ed. International Plant Nutritional Institute, South East Asia Program. Yield gap analysis with local to global relevance—a review. F. Crop. Res. 143:4–17.
FAOSTAT, 2015. Database. http://faostat3.fao.org/home/E. http://dx.doi.org/10.1016/j.fcr.2012.09.009.
Fischer, T., Byerlee, D., Edmeades, G., 2014. Crop Yields and Global Food Security: Will Von Uexküll, H., Henson, I., Fairhurst, T., 2003. Canopy management to optimize yield. In:
Yield Increase Continue to Feed the World? ACIAR Monograph No. 158. Australian Fairhurst, T., Härdter, R. (Eds.), Oil Palm: Management for Large and Sustainable
Centre for International Agricultural Research, Canberra. Yields, pp. 163–180.