Controls of water and energy fluxes in oil palm plantations: Environmental variables and oil palm age

Agricultural and Forest Meteorology 239 (2017) 71–85

Contents lists available at ScienceDirect

Agricultural and Forest Meteorology journal homepage: www.elsevier.com/locate/agrformet

Controls of water and energy fluxes in oil palm plantations: Environmental variables and oil palm age Ana Meijide a,∗ , Alexander Röll b , Yuanchao Fan a , Mathias Herbst c , Furong Niu b , Frank Tiedemann a , Tania June d , Abdul Rauf e , Dirk Hölscher b , Alexander Knohl a a

University of Goettingen, Bioclimatology, Büsgenweg 2, 37077, Göttingen, Germany University of Goettingen, Tropical Silviculture and Forest Ecology, Büsgenweg 1, 37077, Göttingen, Germany c Thünen Institute of Climate-Smart Agriculture, Bundesallee 50, 38116 Braunschweig, Germany d Bogor Agricultural University (IPB), Department of Geophysics and Meteorology, Bogor, Indonesia e University of Tadulako (UNTAD), Faculty of Agriculture, Palu, Indonesia b

a r t i c l e

i n f o

Article history: Received 24 June 2016 Received in revised form 11 February 2017 Accepted 26 February 2017 Keywords: Eddy covariance Sap flux Albedo Transpiration Energy fluxes Evapotranspiration

∗ Corresponding author. E-mail address: [email protected] (A. Meijide). http://dx.doi.org/10.1016/j.agrformet.2017.02.034 0168-1923/© 2017 Elsevier B.V. All rights reserved.

a b s t r a c t Oil palm is rapidly expanding, particularly in Indonesia, but there is still very limited information on water and energy fluxes in oil palm plantations, and on how those are affected by varying environmental conditions or plantation age. In our study, we measured turbulent fluxes of sensible (H) and latent (LE) heat and gross primary productivity (GPP) with the eddy covariance technique for 8 months each in a young oil palm plantation (1-year old) and subsequently in a mature plantation (12-year old) in Jambi Province, Sumatra, Indonesia. Simultaneous measurements of transpiration (T) were performed using the sap flux technique. We additionally estimated albedo, the maximum rate of carboxylation (Vcmax ), the maximum rate of photosynthetic electron transport (Jmax ) and water use efficiency (WUE). LE dominated the energy budget in both plantations, particularly in the mature one, where it accounted for up to 70% of the available energy. In the young oil palm plantation, evapotranspiration (ET) was significantly reduced and H fluxes were higher. The Bowen ratio was higher in the 1-year old plantation (0.67 ± 0.33), where it remained constant during the day, than in the mature plantation (0.14 ± 0.09), where it varied considerably over the day, suggesting the existence of water sources inside the canopy which evaporated during the day. Albedo was similar in both plantations (0.16 ± 0.02 and 0.14 ± 0.01 for the 1 and 12-year old plantation, respectively), while WUE differed with plantation age. Annual T estimates for oil palm were 64 ± 3 and 826 ± 34 mm yr−1 for the 1 and 12-year old plantation, respectively. The corresponding annual ET was 918 ± 46 and 1216 ± 34 mm yr−1 , respectively. The Community Land Model (CLM), a process based land surface model that has been adapted to oil palm functional traits (i.e. CLM-Palm), was used to investigate the contribution of different water sources to the measured fluxes. CLM-Palm differentiates leaf and stem surfaces in modelling water interception and thus is able to diagnose the fraction of dry foliage that contributes to T and the wet fraction of all vegetation surfaces (leaf and stem) that contributes to evaporation. The results of the simulations performed are consistent with the storage of water within the canopy in the mature plantation, and suggest that oil palm trunk surfaces including epiphytes provide water reservoirs for intercepted rain which significantly contribute to ET. The decoupling between GPP and T in the morning and the early decreases of both fluxes at midday point to internal water storage mechanisms in oil palms both in the leaves and in the stem, which delayed the detection of water movement at the leaf petioles. Our measured data combined with the model simulations therefore suggest the existence of both external and internal trunk water storage mechanisms in mature oil palms contributing to ecosystem water fluxes. Oil palm plantations can lead to surface warming at early stages of development, but further assessments should be performed at landscape level. Our study provides data relevant for the parametrization of larger-scale models, which can contribute to understanding the climatic feedbacks of oil palm expansion. © 2017 Elsevier B.V. All rights reserved.

72

A. Meijide et al. / Agricultural and Forest Meteorology 239 (2017) 71–85

1. Introduction The rapid demand growth for vegetable oils is supporting the expansion of oil palm (Elaeis guineensis Jacq.) plantations, particularly in Indonesia, which are currently responsible for about half of the world’s palm oil production (OECD-FAO, 2012; FAO, 2012). The surface dedicated to oil palm cultivation in Indonesia has increased at a rate of 400 000 ha yr−1 from 2008, reaching a total cultivated area of 10.6 million ha in 2015 (USDA, 2016). The existing production is concentrated in Sumatra (OECD-FAO, 2012), where large land-use changes have taken place, e.g. losses of over 21% of forest cover between 2000 and 2012 in the Sumatran lowlands (Margono et al., 2014). Oil palm expansion affects different ecosystem properties and functions such as biodiversity (Barnes et al., 2014), soil carbon storage (Guillaume et al., 2015), biomass carbon pools (Kotowska et al., 2015) and greenhouse gas cycling (Carlson et al., 2012). A recent study in Borneo found up to 6.5 ◦ C higher air temperatures in oil palm plantations than in primary forests (Hardwick et al., 2015), indicating that the conversion of forest to oil palm plantations and associated changes in canopy coverage can also have severe micro-climatic effects by modulating the land-atmosphere fluxes of energy and water (Alkama and Cescatti, 2016). Forest conversion to oil palm was also reported to impact on water cycling including evapotranspiration (ET) and infiltration rates (Comte et al., 2012; Merten et al., 2016). Understanding the underlying mechanisms of such changes requires a comprehensive assessment of ecosystem energy fluxes in oil palm plantations, particularly the partitioning of available energy into ET (latent heat flux, LE) and sensible heat flux (H). Changes in their magnitude or their ratio could for example, affect regional weather and climate patterns (Aragao, 2012; Pielke et al., 1998). Climate feedbacks due to land conversion have been observed in other regions (Bounoua et al., 2002; Defries et al., 2002), showing changes in albedo (Betts, 2000; Schwaiger and Bird, 2010) or increased temperature in the tropics and subtropics (Bounoua et al., 2002). Conversion to oil palm plantations could trigger thus far unassessed climatic feedback mechanisms. Regarding changes in the water cycle after conversion to oil palm, some hydrological studies showed changes mainly during the first years after land clearing, which subsequently seemed to dissipate in mature oil palm plantations (Comte et al., 2012). Attempts to assess the effects of oil palm age on some of the most relevant ecosystem water fluxes, i.e. transpiration (T) and ET, indicated large increases of T with increasing oil palm age during the first 10 years of cultivation, and ET rates increased 1.7 fold from young to mature oil palm plantations (Röll et al., 2015). Microclimatic effects, on the other hand, seem to be strongest during the early stages of oil palm development (Luskin and Potts, 2011). Consequently, for a comprehensive assessment of the effects of oil palm expansion on energy fluxes, changes in their partitioning need to be evaluated at different stages of plantation development. Young oil palms, during the first years after plantation establishment, do not produce fruits. After 2–3 years, the palms mature, and fruits can be harvested (Dislich et al., 2016), making plantations productive in terms of yield. Evapotranspiration and energy partitioning are currently being studied across a variety of ecosystems using the eddy covariance (EC) technique as part of an international network of flux measurements (Aubinet et al., 1999; Running et al., 1999). The EC technique allows for simultaneous quantification of LE, H and carbon dioxide (CO2 ) fluxes, and, in combination with meteorological measurements, can provide information on the environmental and physiological controls of the studied fluxes. Only a few studies using the EC technique have so far been carried out in oil palm plantations, and they were limited to very short periods of time (Röll et al., 2015; Fowler et al., 2011; Henson and Harun, 2005). They therefore

provide only limited information on the variation of these fluxes under changing environmental conditions. Ecosystem carbon and water cycles are intimately coupled by gas exchange through plant stomatal conductance (gs ), which regulates photosynthesis and T rates (Collatz et al., 1991). Previous studies on oil palms have shown that T peaks early in the morning, which may be facilitated by internal trunk water storage mechanisms (Röll et al., 2015). Early peaks followed by subsequent declines of oil palm T indicate stomatal closure, which not only affects water fluxes (ie. T and ET), but also CO2 uptake. The analysis of canopy conductance (gc ) derived from EC measurements together with leaf measurements of gs can provide insight on some of the underlying physiological controls of stomatal closure and its impact on water and carbon fluxes. A further relevant ecosystem indicator of carbon-water interaction is the water use efficiency (WUE), i.e. the ratio of carbon uptake during CO2 assimilation to water loss during T (Ehleringer, 1993). It is a major factor for the survival, productivity and fitness of plants (Osmond et al., 1982) and an indicator of their resilience to changing climatic conditions (Chaves, 2004). Comparative studies of WUE can give insight on how future climatic changes may affect ecosystems carbon and energy budgets (Ponton et al., 2006). For oil palms, it is expected that ecosystem scale WUE varies substantially between highly productive and non-productive plantations. In order to predict long-term weather and climate changes associated to the rapid expansion of oil palm plantations in many tropical areas, large-scale models are required to upscale effects from the stand and watershed level to the global scales (Granier et al., 2000). However, parametrization of such large-scale models should be informed by field measurements (see Tenhunen et al., 1998), but despite the extent of oil palm expansion and associated deforestation and land-use change, information on key eco-physiological variables is still relatively scarce. For maritime conditions such as in Indonesia, climatic impacts due to land-use change are expected to be even stronger than under continental conditions, as 40% of the global tropical latent heating of the upper troposphere takes place over the Maritime Continent (van der Molen et al., 2006). Thus, there is a need for field measurement of essential variables such as the photosynthetic capacity of oil palms (i.e. Vcmax , the maximum rate of carboxylation, and Jmax , the maximum rate of photosynthetic electron transport) and crop coefficients to estimate ET (Allen et al., 1998) derived from field measurements. One land-surface model that allows for a systematic process-based simulation of carbon, water and energy exchanges between land and atmosphere as well as of microclimatic effects is the Community Land Model (CLM, Oleson et al., 2013). Recently, a CLM-Palm sub-model (Fan et al., 2015) has been developed to simulate palm species within the framework of the CLM4.5. It can therefore be used to facilitate the understanding of the underlying mechanisms of water and energy flux regulation and partitioning. The parameterization of canopy hydrology is critical for modelling water fluxes and for partitioning the available energy into H and LE. However, there is a lack of understanding of the fraction of intercepted precipitation and its effects on leaf gas exchange among the major land-surface models and even among different versions of the CLM (De Kauwe et al., 2013; Lawrence et al., 2007). Our research question in this study was understanding energy fluxes and their environmental controls in oil palm plantations and how they differ with plantation age. Therefore, the objectives were 1) to quantify oil palm characteristics relevant for water and energy exchange, e.g. albedo, WUE, crop coefficients; 2) to evaluate water and energy fluxes at different stages of oil palm development; and 3) to identify the controlling physiological mechanisms and environmental factors of the studied fluxes. To achieve these objectives, we carried out EC measurements in a 1-year old oil palm plantation in the Jambi Province, Sumatra, Indonesia for 8 months.

A. Meijide et al. / Agricultural and Forest Meteorology 239 (2017) 71–85

Measurements were subsequently performed in a nearby 12-year old plantation (also for 8 months). In conjunction with the EC measurements, T was derived from sap flux density measurements with thermal dissipation probes (TDP; Granier, 1985). Simultaneously, a variety of environmental variables were recorded. They serve to investigate main drivers of temporal variability of energy and water fluxes as well as to parametrize stand-scale models for deriving yearly estimates of water consumption by oil palms. Additionally, gs Vcmax and Jmax were measured at the leaf level and related to potential driving variables, while the CLM-Palm was used to simulate the underlying mechanisms for the partitioning of the studied fluxes. 2. Material and methods 2.1. Study sites In our study we compare two sites in the Jambi Province, Sumatra, Indonesia, a 1-year old and a 12-year old oil palm plantation. The sites are located 19 km apart and have similar climatic conditions. Between 1991 and 2011, average annual temperature in the area of study was 26.7 ± 0.2 ◦ C (1991–2011 mean ± SD), with little intra-annual variation (Drescher et al., 2016). Annual precipitation in the same period was 2235 ± 385 mm and a dry season with less than 120 mm monthly precipitation usually occurred between June and September (temperature and rainfall data from Airport Sultan Thaha in Jambi, located at approximately 28 and 45 km from the study sites, Drescher et al., 2016). Soils in the area are loam Acrisols (Allen et al., 2015; Guillaume et al., 2015). Soil carbon contents for the 1 and 12-year old sites were 1.11 ± 0.48 and 1.12 ± 0.34%, respectively (mean ± SD); respective nitrogen contents were 0.08 ± 0.03 and 0.08 ± 0.02% (mean ± SD). The terrain is generally flat with small elevation variations (+/− 15 m). The 1-year old oil palm plantation (1◦ 50 7.6”S, 103◦ 17 44.2”E, 75 m a.s.l), from here on referred to as Pompa Air, was a small holder plantation of approximately 5.7 ha. Palms were on average 2 m high during the period of measurements and still did not produce fruits. Over 60% of the soil was covered with grasses and seasonal crops (Röll et al., 2015), e.g. cassava (Manihot esculenta), which were grown between the oil palms and were cut every 3–4 months. The 12-year old plantation (1◦ 41 35.0”S, 103◦ 23 29.0”E, 76 m a.sl.) was property of a big company (PTPN VI) with a plantation area in the location of study of 2186 ha. Canopy height was 12 m, LAI was 3.64 m2 m−2 (Fan et al., 2015) and the oil palms were mature and therefore produced fruits, with a mean annual yield during the years of study (2014–2015) of 27.7 Mg ha−1 (first harvests commonly 3–5 years after planting). The plantation was fertilized in the years of study with up to 196 kg N ha−1 applied as urea. There was hardly any understory vegetation (grasses), as herbicides were routinely sprayed, but butts of pruned oil palm leaves along the trunks were densely covered with epiphytes. Measurements were carried out from July 2013 to February 2014 in the 1-year old plantation and from May 2014 to February 2015 in the 12-year old plantation (Appendix Table A1). 2.2. Eddy covariance fluxes At both sites, the EC technique was used to derive LE, H and net ecosystem CO2 exchange (NEE) from high frequency (10 Hz) measurements of CO2 and water vapor (H2 O) concentration and air temperature above the canopy. The flux system consisted of a sonic anemometer (Metek uSonic-3 Scientific, Elmshorn, Germany) and a fast response open-path CO2 /H2 O infrared gas analyzer (Li-Cor 7500A, LI-COR Inc. Lincoln, USA). Digital outputs from both instruments were recorded in a computer program by a custom based

73

ProfiLab-Expert (Abacom, Ganderkeese, Germany) (Laurila et al., 2005). They were subsequently aligned following an approach similar to Eugster and Plüss (2010), where the sonic anemometer data are used as reference and the most proximate Li-Cor 7500A data is matched. In this study, LE, H and GPP fluxes follow conventional use (Chapin et al., 2006), i.e. they are positive when directed away from the surface. We used a 7 m tower at the 1-year old plantation and a 22 m tower at the 12-year old plantation. Instruments were placed in a boom located at the top of both towers at 7.4 and 22.4 m respectively. Fluxes were calculated with the software EddyPro (LI-COR, Lincoln, USA), planar-fit coordinate rotated (Wilczak et al., 2001), block-averaged and corrected according to the Webb-PearmanLeuning (WPL) theory to compensate for air density fluctuations (Webb et al., 1980). Spectral corrections were performed according to Moncrieff et al. (2005, 1997). Thirty-minute flux data were flagged for quality applying the steady state and integral turbulence characteristic test (Mauder and Foken, 2006). Three categories were established, and data belonging to class 2 (steady state deviation >100% or integral turbulence characteristics deviation >100%) were discarded. Storage fluxes were estimated from single point measurements at the top of the tower. Energy balance closure for all 30 min data during the measurement period was 73 and 84% for the 1 and the 12-year old plantation, respectively. Data were also filtered according to friction velocity (u*) to avoid the possible underestimation of fluxes in stable atmospheric conditions (Goulden et al., 1996); u* was derived for each site following (Reichstein et al., 2005) and u* threshold of 0.1 m s−1 was found for both sites. Footprint analyses were performed for both locations following Kljun et al. (2004), i.e. data points with an anemometer-to-flux distance exceeding the plantation limits (assessed separately in 12 segments of 30◦ ) were excluded from the analysis The 1-year old plantation had a polygonal asymmetrical shape, with a mean plantation size in the 12 segments of 114.8m, and a minimum and maximum of 78 and 204 m, respectively. When the fetch was greater than the plantation size in any of these directions, data points were removed from our analysis. For the 12-year old site, the same procedure was applied, but only data coming from wind directions between 255–345◦ were considered, as the plantation extended over several kilometers in all other directions. Tower distance to the end of plantation in the evaluated areas was 763–823 m. Footprint analysis resulted in discarding 3% and <1% of available data for Pompa Air and PTPN VI, respectively. The number of data points removing on each of the filtering processes for both sites can be seen on Appendix Table A2. Gap-filling of meteorological and eddy covariance data was carried out similarly to (Falge et al., 2001), but considering the co-variation of fluxes with meteorological variables and the temporal auto-correlation of the fluxes (Reichstein et al., 2005). NEE data were partitioned according to Reichstein et al. (2005) to derive gross primary productivity (GPP). 2.3. Sap flux measurements Sap flux measurements were carried out at both sites as explained in Röll et al. (2015), following a methodological approach for sap flux measurements on oil palms (Niu et al., 2015). Thermal dissipation probe (TDP) sensors were installed in the leaf petioles of 16 leaves at both sites, 4 each on 4 different palms. The sensors were connected to AM16/32 multiplexers connected to CR1000 data loggers (both Campbell Scientific Inc., Logan, USA). The signals from the sensors were recorded every 30 s and averaged and stored every 10 min. The mV-data from the logger were converted to sap flux density with the empirically-derived calibration equation by Granier (1985), but with a set of equation parameters that

74

A. Meijide et al. / Agricultural and Forest Meteorology 239 (2017) 71–85

was specifically derived for TDP measurements on oil palm leaf petioles (Niu et al., 2015). To ensure that nighttime zero-flux conditions were met in the two plantations, we examined the nighttime sap flux values for each sensor during the measurement period with respect to the values adjacent to the respective maximum daily differential temperature (i.e. Tmax, following Oishi et al., 2008). They remained stable over several hours during the early morning hours, when vapor pressure deficit (VPD) was consistently below 0.1 kPa; nighttime zero-flux conditions as a precondition for accurate estimates with the thermal dissipation probe method were thus met in our study. Individual leaf water use rates (e.g. aggregated to hourly, daily) were calculated by multiplying respective sap flux densities by the water conductive areas of the leaves; the latter were derived from a linear regression with leaf baseline length established by Niu et al. (2015) for oil palm petioles in the research region. The water use rates of all individual leaves measured simultaneously (a minimum of 13 leaves running in parallel out of the 16 measured leaves per site) were averaged. To scale up to average palm water use, average leaf water use rates were multiplied by the average number of leaves per palm (counted on 4 palms). Multiplying the average palm water use by the number of palms per unit of land yielded stand T rates. This approach is associated with sample size related estimation errors of stand transpiration of 14% (Niu et al., 2015). 2.4. Environmental variables At both sites, meteorological variables were measured every 15 s and then averaged and stored in a DL16 Pro data logger (Thies Clima, Göttingen, Germany) as 10 min mean, minimum and maximum values. Upwelling and reflected photosynthetic active radiation (PAR) were measured with PQS1-PARQuantum sensors (Kipp & Zonnen, Delft, The Netherlands), and global (Rg) and diffuse radiation with a sunshine sensor (BF5, Delta T, Cambridge, UK). Net radiation (Rnet) and its components were measured with a CNR4 Net radiometer (Kipp & Zonnen, Delft, The Netherlands). All radiation measurements took place above the canopy. Air temperature and humidity were measured with thermohygrometers (type 1.1025.55.000, Thies Clima, Göttingen, Germany), at four heights in the 1-year old plantation (2.4, 3.2, 5.8 and 6.8 m) and at six heights in the 12-year old plantation (0.9, 2.3, 8.1, 12.3, 16.3 and 21.7 m). A wind direction sensor (Thies Clima, Göttingen, Germany) was also installed on both towers, together with 3-cup anemometers (Thies Clima, Göttingen, Germany) at 2.4, 3.2, 4.1 and 5.8 m at Pompa Air and 2.3, 13, 15.4 and 18.5 m at PTPN VI, to derive wind velocity profiles. Two precipitation gauges (Thies Clima, Göttingen, Germany) recorded rainfall above the canopy at both towers. Albedo was estimated as the ratio between the reflected and the incoming short wave radiation using data where Rg exceeded 100 W m−2 and which were recorded between 10:00 and 14:00 h as to avoid low sun angles (Domingo et al., 2000; Li et al., 2006). We measured soil heat flux (G) with three heat flux plates (model HFP01, Hukseflux, Delft, The Netherlands) installed 3 and 5–6 cm below ground at Pompa Air and PTPN VI respectively. Additionally, three soil temperature and soil moisture profiles were measured with soil moisture sensors (Trime-Pico 32, Imko, Ettlingen, Germany) installed at 0.3, 0.6 and 1 m depth. An additional meteorological station, from now on called Reki, located at 33 and 47 km from Pompa Air and PTPN VI, respectively, continuously measured air temperature and humidity, wind speed and rainfall (using the same type of instruments as at the EC towers), as well as global radiation with a CMP3 pyranometer (Kipp & Zonen, Delft, The Netherlands). Data were also recorded with a DL16 Pro data logger every 10 min. This meteorological station was running for one full year without any data gaps, overlapping with

both measurement periods at the EC towers. Data measured at this station were thus used for modelling annual ET and T rates.

2.5. Modelling of annual evapotranspiration and transpiration rates and calculation of water-use efficiency We used the FAO Penman-Monteith equation (FAO 56: Allen et al., 1998) to derive crop coefficients based on measured data by comparing measured and potential ET (ET EC and ET pot, respectively). Crop coefficients for the young and the mature oil palm plantations were derived from the slope of the linear regression between measured ET EC and modeled ET pot daily values. The full meteorological data series from the Reki station (uninterrupted measurements for one year) allowed for calculating daily and yearly ET pot rates. Subsequently, to get yearly ET values for the two plantations, we multiplied the respective crop coefficients by the yearly ET pot values derived from meteorological measurements at the Reki station. We also used the full yearly data series of daily ET pot values from the Reki station to derive estimates of yearly oil palm T for the two sites. In analogy to deriving ET from ET pot, we took the slopes of the linear regression lines (forced through origin) between measured daily T (n = 51 days at Pompa Air, 66 days at PTPN VI) and calculated ET pot from the Reki station to estimate daily and thus yearly T rates for the two plantations. To provide uncertainty estimates for the yearly modeled values of ET and T, we performed parametric bootstrapping (with the R package ‘boot’) with 50000 repetitions and derived means and confidence intervals from the resulting distributions. For ET, we bootstrapped the linear relationship between ET pot and ET EC (regressions forced through origin), i.e. the crop coefficients. For T, we bootstrapped the linear relationship between ET pot and T (regressions forced through origin). The resulting corridors (confidence intervals) from the bootstrapping could subsequently be used to provide a measure of the uncertainty of annual ET and T estimates. Water use efficiency at the ecosystem scale (WUE) was calculated as the ratio of gross primary productivity (GPP, g C m−2 day−1 ) to ET (Reichstein et al., 2005).

2.6. Measurements of stomatal conductance (gs ), maximum velocity of carboxylation (Vcmax ), maximum electron transport rate (Jmax ) and estimations of canopy conductance (gc ) Stomatal conductance (gs ) was measured at PTPN VI and in a 1-year old oil palm plantation in 2015, two years after the EC measurements at Pompa Air. We did not measure gs at the original Pompa Air site, as it had turned productive in the meanwhile. We instead rather chose a nearby non-productive plantation comparable to the former conditions at Pompa Air (at 3 km distance). Measurements were performed using a leaf porometer (SC1, Decagon Services, Pullman, USA) on 3 consecutive days every 30–60 min (between 7 am and 6pm), on 3 different palms and on leaves at 3 different heights in each of the plantations. Stomata in oil palm are mainly in the abaxial surface (Luis et al., 2010) and therefore we measured on this surface. Velocity of carboxylation (Vcmax ) and Jmax were also measured in both sites using a Licor 6400 (Licor Inc., Licoln, USA), on 3 palms in each of the plantations and selecting petioles at 3 different heights in each of the palms. Data were processed using the R package Plantecophys (Duursma, 2015). Canopy conductance (gc , m s−1 , converted to mol m−2 s−1 ) was approximated from surface conductance (gf , m s−1 , converted to mol m−2 s−1 ) as calculated with the reversed Penman-Monteith

A. Meijide et al. / Agricultural and Forest Meteorology 239 (2017) 71–85

equation (e.g., Grace et al., 1995) as explained in Knohl and Buchmann, (2005): 1 = 1.6gf

sH g  a

+ CpD LE

−

1 ga

(1)

where s represents the slope of the saturation vapor pressure vs. the temperature curve (kg kg−1 K−1 ), H is the sensible heat flux (W m−2 ), LE is the latent heat flux (W m−2 ), Cp represents the heat capacity of air (J kg−1 K−1 ), D is the vapor pressure deficit (kg kg−1 ) and ␥ represents the psychometric constant (kPa K−1 ). As H and LE had the same footprint, H was used instead of an energy balance term for H (Shuttleworth et al., 1984). Aerodynamic conductance for CO2 (ga ), water vapor (g a ) and heat (g a ), respectively (all in m s−1 ), were calculated as: 1 1 1 = + ga gt gb

(2)

1 1 1 = + ga gt 1.4gb

(3)

1 g  a

=

1 gt

momentum



1 (4)with the transfer conductance for 1.4∗0.92gb ¯ m s−1 ) and the excess conductance (gt = u∗2 /u, −2 −1 3

+



(g b = u ∗ B Sc/Pr

,ms

), where u* denotes friction velocity

¯ (m s−1 ) and uthe mean horizontal wind speed (m s−1 ), B is the inverse Stanton number (B = 1/5 after Kramm et al., 2002), Sc is the Schmidt number for CO2 (Sc = 1.02), and Pr is the turbulent Prandtl number (Pr = 0.72). For gc diurnal patterns, we used medians of all available data for dry periods (i.e. no rainfall in the previous 3 days). 2.7. Flux simulations using CLM-Palm To simulate carbon and energy fluxes, we used CLM-Palm (Fan et al., 2015), a sub-model developed for palms within the frame of the CLM4.5 (Oleson et al., 2013). It was previously calibrated and validated for site conditions, growth and yield of oil palms in our study region (Fan et al., 2015), including the 12-year old plantation investigated in this study. CLM-Palm only simulates a single plant functional type (PFT) in a vertical column. Therefore, the collective fluxes in the 1-year old plantation, in which large amounts of grasses and crops were grown in between the oil palms, could not be simulated in a comparable way to the EC data. Thus, only measured fluxes from the 12-year old plantation, where the undercanopy vegetation was minimal, were compared to fluxes derived with the CLM-Palm model. The CLM4.5 uses two fixed parameters controlling canopy interception of precipitation (Lawrence et al., 2007). The first is related to the water interception efficiency by foliage and stem surfaces (fpimx), which represents the fraction of precipitation intercepted by the canopy and comprises both the fraction that evaporates back into the atmosphere and the water that is then transported through the stem as stemflow (and 1-fpi is the fraction of direct throughfall). The second parameter is the maximum water film thickness that can be stored on these surfaces (dewmx). The exceeding water is partitioned to canopy runoff. Both the intercepted water (fpi) and dewmx are updated based on canopy interception gains and evaporation losses. Details on the CLM canopy hydrological parameterization are reported in Fan (2016). The parameters fpimx and dewmx should vary depending on canopy morphological traits. In the current CLM, these parameters are constant for all plant functional types, from natural forests to crops. An apparent weakness of this uniform parameterization is that potential differences in the interception efficiency and capacity of different plants are not accounted for. In the first versions of CLM (CLM3, Oleson et al., 2004), the original value for fpimx was

75

1.0. Nevertheless, Lawrence et al. (2007) reduced this value to 0.25 in order to obtain a better fitting of the partitioning of ET (by reducing the proportion of canopy evaporation) to global mean patterns. While rainforests are characterized by an extremely dense canopy and diverse leaf morphology, oil palm plantations have incomplete canopy cover and the palms have uniform, large leaves with caved leaflets and leaf axils on the trunk (Carr, 2011). The latter are assumed to be able to capture and store significant amounts of precipitated water (Merten et al., 2016). Biases in canopy water interception could directly propagate to T and ET calculations in CLM because the intercepted precipitation will decrease T by reducing the fraction of dry canopy that is available for T (Lawrence et al., 2007). A preliminary comparison of the default model with the EC flux data of the mature oil palm plantation PTPN VI showed a substantial underestimation of ET and LE and an overestimation of T and H by the model (see Fig. 6 and Appendix Table A3). One potential explanation could be that CLM under-estimates the canopy water interception by oil palms. An increase in the intercepted precipitation by the canopy could lead to higher evaporative fluxes (thus lower T to ET ratio) from the wet leaf surfaces and water reservoirs accumulated in the axils of pruned leaves along the stem. This underestimation would take place especially around noon, when radiation, VPD and temperatures are near their maximum daily values. An additional hypothesis for model divergences could be that CLM underestimates soil evaporation. However, due to field observations of water storage in leaf axils along the trunk, which had previously been observed by Merten et al. (2016), we focused our model experiments on evaluating the canopy processes. Based on the initial model evaluation and field observation, we hypothesize that biases in modelling water fluxes (ET, T) and energy partitioning (LE, H) for the oil palm plantation could be reduced by adapting the CLM canopy hydrological scheme to oil palm canopy traits through the following aspects: 1) the interception efficiency parameter (fpimx) of CLM4.5 should be increased to the original value in CLM3 to better fit oil palm’s canopy structure; 2) oil palm leaf surfaces might be able to hold substantially thicker water films (dewmx) than the default value (0.1 mm); 3) the water storage capacity of oil palm stem surfaces should be separately modeled because of the formation of water reservoirs in leaf axils. A series of model experiments were conducted to test the above hypothesis, e.g. successively increasing fpimx to the original value in CLM3 (from 0.25 to 1.0) and increasing dewmx (from 0.1 to 0.4) to values comparable to those used in other models (i.e. Canoak, Baldocchi, 1997). Further model adaptations focused on incorporating the different water storage capacities of leaflets, rachis and axils into the CLM-Palm model. The rachis and axils were modeled as stems (and thus part of the stem area index, SAI). Two dewmx parameters, dewmx1 and dewmx2, were thus set for the leaves (LAI) and stems (SAI), respectively, according to field observations showing that stem surfaces are able to hold a much thicker water film than leaf surfaces (up to 6 mm, Merten et al., 2016). Based on this differentiation, the total intercepted water could be partitioned between leaf and stem surfaces and the fractions of wet vs. dry canopy could be calculated. This was an essential further step, as only the dry fraction of foliage (fdry of LAI) is considered relevant for T and photosynthesis in the CLM, whereas the wet fraction of all vegetation surfaces (LAI + SAI) is considered when modelling evaporation. A detailed description of this new canopy hydrological parameterization can be found in Fan (2016). All statistical analyses and data filtering were performed with R studio version 3.1.1 (R Core Team, 2014); for graphing, Origin 8.5 (Origin Lab, Northampton, MA, USA) was used.

1251 mm (232 days) 28.2 ± 1.7 826 ± 34 1216 ± 34 1.06 (1.03, 1.09) 88.96 ± 3.45 42.00 ± 2.42 0.14 ± 0.09 0.14 ± 0.01

2.76 ± 0.40

1834 mm (230 days) 26.3 ± 1.8 64 ± 3 918 ± 46 0.80 (0.77, 0.84) 81.47 ± 6.59 51.97 ± 9.43 0.67 ± 0.33

3.29 ± 0.32

Mean soil moisture at 0.3 m during measurement period vol (%) (mean ± SD) Annual T (mm yr−1 ) Annual ET (mm yr−1 ) Crop coefficient FAO-PM (daily) Jmax (␮mol CO2 m−2 s−1 ) Vcmax (␮mol CO2 m−2 s−1 )

0.15 ± 0.02

Latent heat flux was higher than H in both of the studied plantations, but the partitioning of energy fluxes into H and LE was strongly influenced by plantation age. Net radiation was more evenly partionioned in the 1-year old Pompa Air plantation (55% and 34% for LE and H at midday, respectively) than in the 12-year old PTPN VI plantation, where most of the energy was used for latent heat fluxes (78% and 12% for LE and H at midday, mean diurnal cycles, Fig. 2a and b). At the 1-year old site, the diurnal patterns of H and LE were similar, i.e. the ratio between sensible and latent heat flux (Bowen ratio, ␤) remained relatively constant (at about 0.60) throughout the daytime hours, particularly from 9 to 16 h. At the 12-year old plantation, ␤ was diurnally highly variable and decreased from 0.27 to 0 during the same time interval. Midday ␤ for both plantations during the whole measurement period are presented in Table 1. At both plantations, H and LE started to increase during the first hours of the day at around 6–7 h. At the 1-year old site, they

1-year old (Pompa Air) 12-year old (PTPN VI)

3.2. Diurnal patterns and partitioning of energy fluxes

WUE (g CO2 kg H2 O−1 ) (mean ± SD)

Total rainfall exceeded 1200 mm in both sites during the 8 months measurement period and soil moisture at 0.3 m was never below 20% in any of the sites, with mean values during the whole measurement period of 26.3 ± 1.8 and 28.2 ± 1.7% for the 1 and 12-year old plantations respectively (Table 1). Midday albedo was very similar at both sites, (0.15 ± 0.02 and 0.14 ± 0.01 for Pompa Air and PTPN VI, respectively, Table 1), and no significant variations were observed during the 8 months measurement period. Water use efficiencies for the whole ecosystem under dry conditions (calculated for the whole measurement period for each site on dry days, i.e. three consecutive days without rainfall) as the ratio of daily GPP to daily ET were 3.29 ± 0.32 g C kg−1 H2 O (mean ± SD) in the 1-year old and 2.76 ± 0.40 g C kg−1 H2 O in the 12-year old plantation (Table 1). On the days chosen for the analysis (i.e. under dry conditions), WUE was not significantly influenced by day-to-day variations in radiation or VPD. Vcmax (the maximum rate of carboxylation) was higher in the 1-year old plantation (51.9 7± 9.43 ␮mol CO2 m−2 s−1 , see Table 1) than in the 12-year old one (42.00 ± 2.42 ␮mol CO2 m−2 s−1 ), while both plantations had similar maximum rates of photosynthetic electron transport, Jmax (81–89 ␮mol CO2 m−2 s−1 , Table 1). We derived crop coefficients for the commonly used FAO Penman-Monteith equation (Allen et al., 1998) for the 1-year old and the 12-year old plantation by comparing daily measured ET (ET EC) with potential ET (ET pot, derived following Allen et al. (1998) using the environmental variables measured at the Reki station). There was a strong linear relationship (R2 ≥ 0.99, P < 0.001) between ET EC and ET pot for both sites (Fig. 1). Respective crop coefficients for the 1 and the 12-year old plantations derived from the regression equations were 0.8 (Pompa Air) and 1.06 (PTPN VI) (Table 1). Using these crop coefficients and the ET pot values derived from available climatic yearly data series from the Reki meteorological station allowed to calculate annual ET for the two plantations. Annual ET was 918 ± 46 mm yr−1 for the 1-year old and 1216 ± 34 mm yr−1 for the 12-year old oil palm plantation (Table 1). Annual T rates and the contribution of T to ET, derived from linear regression analysis of T to ET pot data (in analogy to the approach for ET above), were very low in the 2 year old plantation (64 ± 3 mm yr−1 or 7% of ET) and much higher in the 12 year old plantation (826 ± 34 mm yr−1 or 68%) (Table 1).

Midday bowen ratio (mean ± SD)

3.1. Environmental conditions, albedo, water use efficiencies and photosynthesis, crop coefficients and yearly water fluxes

Albedo

3. Results

Rainfall during measurement period (mm)

A. Meijide et al. / Agricultural and Forest Meteorology 239 (2017) 71–85 Table 1 Albedo, bowen ratio, water use efficiency (WUE), the maximum rate of carboxylation (Vcmax ) and the maximum rate of photosynthetic electron transport (Jmax ), crop coefficients for the FAO Penman-Monteith equation (FAO-PM), annual evapotranspiration (ET) and annual transpiration (T) derived from annual estimates of potential evapotranspiration (FAO-PM), mean soil moisture at 0.3 m depth for all available data during measurement period and total rainfall for the 1-year old and the 12-year old oil palm plantations, respectively. Rainfall data were obtained from the Reki station, located at 33 and 47 km from Pompa Air and PTPN VI, respectively, which measured continuously without data gaps during the period of study.

76

A. Meijide et al. / Agricultural and Forest Meteorology 239 (2017) 71–85

77

Fig. 1. Comparison of measured evapotranspiration of oil palms (ET EC, eddy covariance derived) and modeled potential evapotranspiration (ET pot, FAO Penman-Monteith derived) in the 1 year-old (a) and 12 year-old (b) oil palm plantations. Regression lines (forced through origin) are presented for daily values of ET; regression slopes represent crop coefficients for the 1 and the 12-year old oil palm plantation, respectively. Dataset used for this analysis had at least 85% full daytime eddy covariance data coverage.

Fig. 2. Mean diurnal cycles (lines) and standard deviations (shaded areas) of sensible heat flux (H), latent heat flux (LE), net radiation (Rnet) soil heat fluxes (G) and Bowen ratio (␤) in the 1 year-old (a) and 12 year-old (b) oil palm plantations and mean diurnal cycles of vapor pressure deficit (VPD), air temperature (Ta), global radiation (Rg) and albedo in the 1 year-old (c) and 12 year-old (d) oil palm plantations. All available half-hourly values from the respective measurement periods considered.

reached their maximum around midday and thus coincided with the diurnal peak of radiation. At the 12-year old site, H already peaked at 9:30 h and then remained relatively constant until midday, when LE reached its maximum. Both fluxes subsequently decreased throughout the rest of the day (Fig. 2). Soil heat fluxes were higher in the 1-year old plantation, where they peaked at midday, than in the 12-year one, where the peak was less pronounced and fluxes remained relatively high from 11 to 15 h.

3.3. Diurnal patterns of transpiration vs. GPP and stomatal conductance Diurnal courses of GPP and T showed very similar patterns at both sites (Fig. 3), but GPP started to peak in the morning earlier than T. Respective differences were about 1.5 h at the 1year old plantation and 1 h at the 12-year old one. At Pompa Air, maximum T rates were reached at 12:30 h (average of whole mea-

78

A. Meijide et al. / Agricultural and Forest Meteorology 239 (2017) 71–85

Fig. 3. Mean diurnal cycles of CO2 net gross primary productvity (GPP) from eddy covariance measurements and transpiration (T) from sap flux measurements in the 1 yearold (a) and 12 year-old (b) oil palm plantations. Average (lines) and standard deviations (shaded areas) of all available half-hourly values when both T and eddy covariance measurements were running in parallel.

3.4. Water interception simulations

Fig. 4. Relationship between stomatal conductance (gs ) derived from porometer measurements and leaf vapor pressure deficit in the 1-year old (red dots) and 12year old (black dots) plantations. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

surement period), while maximum GPP was observed at 11:00 h. At PTPN VI, maximum T rates were measured at 11:00 h while GPP peaked at 10:30 h, with both fluxes subsequently decreasing. Evening decreases of T and GPP coincided at PTPN VI, but showed a 30 min time lag of GPP to T at Pompa Air. For both the 1 and 12-year old plantation, large diurnal variations in gs were observed, with a decreasing trend from morning to evening (data not shown). In both stands, an inverse relationship between gs and VPD was observed, with gs decreasing at increasing VPD (Fig. 4). In the 12-year old plantation, relative decreases of gs occurred at lower VPD. Penman-Monteith derived diurnal courses of gc under dry conditions had a similar pattern at Pompa Air and PTPN VI (Fig. 5c), with maxima occurring at early hours of the day. However, gc maxima occurred earlier (6:30 h) and were lower (1.9 mm h−1 ) in the 1-year old than in the 12-year old plantation (with gc above 2 mm h−1 until 8 h). Respective diurnal peaks of gc were much earlier than peaks of ET and T (9:30–13 h, Fig. 5b) and radiation (11:30 h) and VPD (14:30 h) (Fig. 5a). This resulted in large hysteresis in the diurnal gc response to VPD (Fig. 5d), particularly in the 12-year old plantation. Attempts to fit Penman-Monteith gc estimates to the Lohammar equation were not successful.

Initial comparison with field data showed that the default water interception parameters in the CLM4.5 yield substantially biased diurnal ET and T patterns compared to EC and sap flux measurements in the 12-year old plantation (Fig. 6a). The simulated ET and T patterns were of similar magnitude throughout the day, which implies that there is a very minimal evaporative component, whereas the measured data showed significantly lower T than ET (2.6 mm day−1 vs. 4.7 mm day−1 , or a T to ET ratio 54%), especially around midday. Energy partitioning between H and LE was also biased in the model, i.e. LE was overestimated and H was underestimated throughout the day (Fig. 6b). The model experiments (see Appendix Table A3) showed that canopy water interception and storage capacity were indeed underrepresented in the original model. Simulations successively increasing the default dewmx and fpimx values to incorporate potentially increased water interception by the oil palm canopy resulted in better agreement of modeled and measured fluxes. Further significant improvement compared to the CLM4.5 default parameterization was achieved by treating respective canopy water storage by leaves and stem areas (LAI vs. SAI) separately (modified model values, Fig. 6a and b and Appendix Table A3). When setting dewmx2 to 6 mm, 50% of the overall storage capacity was accounted for by the stem surface. This adjustment resulted in modeled T patterns similar to those derived from field measurements (Fig. 6a). The separate treatment of leaf and stem for canopy water storage further also substantially improved the estimations of LE and H fluxes in the afternoon (Fig. 6b). 4. Discussion 4.1. Water use efficiencies, photosynthetic characteristics and albedo Species-specific information for the parametrization of largerscale models is not always available. In the case of oil palm, there is still a substantial lack of studies on many energy and water flux related parameters such as WUE, as well as on their photosynthetic characteristics. In our study, we derived WUE as the ratio of daily water loss (ET) to carbon gain (GPP). It was 3.3 and 2.8 g C kg−1 H2 O for the 1 and the 12-year old plantations, respectively. Similar WUE were reported in a tropical rainforest (3.3 g C kg−1 H2 O, Tan et al., 2015), but higher ones for temperate forests (4.4 g C kg−1 H2 O, Tang et al., 2006; 5.4-8.1 g C kg−1 H2 O, Ponton et al., 2006). Changes in WUE associated with changes in land cover may be central to the global cycles of water, energy and carbon (Keenan et al.,

A. Meijide et al. / Agricultural and Forest Meteorology 239 (2017) 71–85

79

Fig. 5. Diurnal patterns of global radiation (Rg) and vapor pressure deficit (VPD) (a), evapotranspiration (ET) and transpiration (T) (b), as averaged from all available halfhourly values when both ET and T measurements were running, and canopy conductance (gc ) calculated from eddy covariance fluxes of latent heat on rainless days (c) for the 1-year old (red) and 12-year old (black) plantations. Relationship between gc and normalized VPD (d); arrows indicate order of observations, numbers indicate respective times. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Fig. 6. (a). Mean diurnal cycles of evapotranspiration (ET, full lines) and transpiration (T, dashed lines). Measured values (‘measured’, black) compared to values modeled with the default CLM.4.5 settings (‘default’, blue) and to values modeled with modified settings with regard to water interception and storage of the canopy (‘modified’, red). Measured data from periods with simultaneous ET (eddy covariance) and T (sap flux) data. (b). Mean diurnal cycles of latent (LE, full lines) and sensible heat (H, dashed lines) fluxes. Measured values (black) compared to values modeled with the default CLM.4.5 settings (‘default’, blue) and to values modeled with modified settings with regard to water interception and storage of the canopy (‘modified’, red). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

2013); however, to our knowledge no studies are currently available where changes in WUE associated with forest conversion in the tropics, and particularly for conversions to oil palm are assessed. As for WUE, values of Vcmax and Jmax for oil palms have not been previously reported. These variables define the photosyn-

thetic capacity of a certain PFT. Our Vcmax and Jmax values from oil palms are comparable to those presented in other studies on tropical species (Ali et al., 2015; Wullschleger, 1993), which have shown median Vcmax values below 50 ␮mol m−2 s−1 and Jmax around 80 ␮mol m−2 s−1 . Vcmax and Jmax are generally related to

80

A. Meijide et al. / Agricultural and Forest Meteorology 239 (2017) 71–85

nitrogen use efficiency in the leaves (Kattge et al., 2009). Therefore, different values would have been expected for the two plantations as the mature one receives a much larger N fertilization input of up to 196 kg N ha−1 yr−1 . However, our results show comparable Jmax for the young and the mature plantation, and higher Vcmax in the 1-year old one. Changes in Vcmax and Jmax with age have also been observed in other species (Han et al., 2008; Kositsup et al., 2010). Niinemets (2002) suggested that increases in plant height rather than in age might be responsible for decreases in net assimilation capacities, while other studies indicate that these changes could rather be explained by variations in nitrogen use efficiency with increasing age (Escudero and Mediavilla, 2003). Our gs measurements were related to oil palm age, showing higher values for a certain VPD in the oil palm leaves of the young plantation than in leaves at the mature plantation (Fig. 4). Due to a much larger LAI in the mature plantation as compared to the young one, this effect was reversed at the canopy level (i.e. looking at gc instead of gs ): canopy conductance was higher at a given VPD in the 12-year old plantation than in the 1-year old one (Fig. 5d). Our LAI measurements in the mature plantation are comparable to other measurements on oil palms of the same age (3.83 m2 m−2 , Awal et al., 2010), which show exponential increases of LAI with oil palm age. This suggest that LAI for the 1-year old plantation should be below the 0.69 m2 m−2 reported in that study for a 2year old plantation (Awal et al., 2010). A much higher LAI in mature plantations compared to young plantations with similar oil palm planting densities could potentially lead to strong differences in albedo. However, our measurements only showed marginal differences between the 1 and the 12-year old plantation. This is likely due to the fact that the soil between palms was not bare in the young plantation, but rather covered with a variety of grasses and seasonal crops, which may have similar reflective properties as oil palm leaves. As oil palms are perennial crops, we observed no significant variation of albedo at either site during the respective measurement periods. Our derived albedos for oil palm plantations (0.14–0.16) compare to values reported e.g. for agricultural systems (pasture) in Amazonia (0.16–0.20), while values of tropical rainforests are commonly lower (i.e. 0.12–0.13, Bastable et al., 1993; Berbet and Costa, 2003; Culf et al., 1995; Gash and Shuttleworth, 1991). Similar results were reported for the temperate zones, i.e. albedos of 0.11–0.15 for different forest types and of 0.18–0.20 for grassland crops (Hollinger et al., 2010). Our results suggest that conversions of forest to oil palm would induce changes in energy balance due to changes in canopy cover and thus albedo. 4.2. Annual water fluxes We calculated crop coefficients for oil palms using ET pot calculated using the FAO Penman-Monteith equation (Allen et al., 1998) and ET derived from EC measurements. Our crop coefficients for the studied 1 and 12-year old oil palm plantations, 0.8 and 1.06, respectively, fall into the 0.8–1.0 range provided by Carr (2011) for oil palm plantations. The crop coefficient of the mature plantation (1.06) is of about the same magnitude as crop coefficients of other oil crops such as rapeseed or sunflower (1.00–1.16) and of other tropical perennial crops (0.95–1.2) including e.g. banana, cacao, coffee, rubber trees or other palm species (Allen et al., 1998), but is e.g. higher than that observed in a tropical coconut plantation (Roupsard et al., 2006). Combining the derived crop coefficients for both plantations with ET pot data series derived from micrometeorological measurements allowed us to calculate the yearly values. Likewise, we derived yearly T rates from the same micrometeorological data series and using a similar approach. Our estimate of PenmanMonteith derived ET pot for our study region was 1147 mm yr−1 , which is comparable to ET pot values provided for the humid trop-

ics in general (927–1076 mm yr−1 , Schlesinger and Jasechko, 2014). For tropical rainforests, 70 ± 14% of ET pot is typically accounted for by plant transpiration (Schlesinger and Jasechko, 2014) which yields a range of 642–964 mm yr−1 for stand T of tropical rainforests including primary forests. Our yearly T estimate for the 12-year old plantation falls within the upper part of this range (826 ± 34 mm yr−1 ), indicating substantial oil palm transpiration in the mature plantation, which was highly productive and heavily fertilized. Transpiration by the 1-year old oil palms was more than 13-fold lower (64 ± 3 mm yr−1 ) due to the very small size of palms and the corresponding low leaf area and leaf number. A substantial contribution of evaporation and non-oil-palm T in the 1-year old plantation reduced the large observed difference in T between the two studied sites to a less than 2-fold difference in ET, i.e. 918 ± 46 mm yr−1 (or an average 2.5 mm day−1 , 1-year old plantation) and 1216 ± 34 mm yr−1 (3.3 mm day−1 , 12-year old plantation). Both the average daily and the annual ET of the mature plantation are similar to ET rates of oil palms derived with a variety of methods (e.g. EC, catchment based, micrometerologically derived), i.e. 1118–1525 mm yr−1 (see review in Carr 2011) or about 3–5 mm day−1 (see summary in Röll et al., 2015). Our values, as well as the values reported in previous studies of oil palm ET, compare to or even exceed ET rates reported for rainforests in South East Asia (e.g. Tani et al., 2003; Kumagai et al., 2005), again pointing to high water fluxes from oil palm plantations despite e.g. much lower stand densities and biomass per hectare than in tropical forests within the same region (Kotowska et al., 2015).

4.3. Diurnal flux patterns and their regulating mechanisms We found strong differences between the 1 and the 12-year old plantations regarding the magnitude and diurnal pattern of LE and H fluxes. While both peaked at midday in the 1-year old plantation, H peaked much earlier (9:30 h) and LE remained higher in the evening in the 12-year old plantation. This led to an early and constant decrease of ␤ in the 12-year old plantation. Similar results have been observed in wet forest canopies (Humphreys et al., 2003; Stewart, 1977) and indicate the presence of water on the trunk or on leaf surfaces. In the 1-year old plantation, ␤ remained relatively constant throughout the day. LE dominated the energy budget in both plantations, particularly in the mature one, where it represented up on average 70% of the available energy (considering the net radiation minus soil heat flux). In addition to the non-stable partitioning of energy fluxes and particularly in the 12-year old plantation, we observed relatively early diurnal peaks, followed by consistent declines throughout the rest of the day, of gc , GPP and T. Transpiration and GPP are coupled through the stomatal opening. In both of the studied oil palm plantations, GPP started earlier in the morning than T (Fig. 3). In the 1-year old plantation, this could be a consequence of T only being measured on oil palms, while GPP integrates over the whole ecosystem and thus includes the CO2 exchange by both palms and the abundant grasses (>60% ground cover). Likewise, GPP remaining high in the afternoon while T already decreases could be an indication of continuing photosynthesis and thus T in grasses under conditions of lower radiation as opposed to oil palms. In the mature plantation, patterns of the studied fluxes were quite different, i.e. there were indications of decoupling between T and photosynthesis. The soil was relatively bare due to the high amounts of herbicides applied in the mature plantation. Even though the trunks of the mature oil palms were densely covered with epiphytes, the latter do not have enough biomass to contribute significantly to ecosystem GPP. Unlike in the young plantation, the earlier start of GPP in the morning thus did not seem to be caused by photosynthetic activity of any other plants in the mature plantation. It rather seems that oil palms start pho-

A. Meijide et al. / Agricultural and Forest Meteorology 239 (2017) 71–85

tosynthesis without significant water movement at the base of the leaf petioles (where sap flux was measured to derive T). Potentially, this could be attributed to an extraction of water from leaf tissues downstream of the sensors. The leaves in the mature plantation, and thus the leaf petioles, were of substantial dimension, i.e. >5 m in length and with a petiole baseline length >12 cm. The dry weight of oil palm leaves from mature plantations in the study region was 3.8 ± 0.9 kg, 75% of which was accounted for by the petiole and 25% by the leaflets (average of 16 oil palm leaves, Kotowska et al., 2015). Average water content of oil palm leaf petioles was reported to be about 60% of fresh weight (Islam et al., 2000; Heuzé et al., 2015), i.e. the petiole of a mature oil palm leaf contains about 4.3 kg of water, at least 70% of which (3.0 kg) would have been located downstream from our sensors. Oil palm leaf water use in the 12-year old plantation was 560 g h−1 during the morning hours (8–10 h average). Thus, to account for the 1 h time lag between NEE and T in the morning, 19% of the petiole water content would have to be extractable, which seems reasonable, as palm leaf tissue is mainly parenchyma. Once this easily accessible water reservoir is depleted, water starts being pulled into the leaves from the trunk, which is detected by the sap flux sensors about 1 h after photosynthesis starts. Additionally to the water storage in the leaves, there might be indications of an easier access to water for oil palms leaves in the morning hours due to a possible contribution of internal trunk water storage mechanisms in mature palms, i.e. GPP and T showed an asymmetric shape with high values before noon and declining values in the afternoon in the 12-year old plantation. This was not observed in the 1-year old plantation, where trunks had not yet formed. In general, gc can be regulated due to leaf-level water stress, limitations in soil moisture or high VPD (Jones and Higgs, 1989; Stewart, 1977). Decreases in T and GPP early in the day, and the associated decreases in gc (Fig. 5d), were probably not a consequence of high VPD, which did not peak until 14 h (Fig. 5a), nor of soil moisture, which remained above 20% during the whole period of study. The observed early diurnal decreases of T and GPP were thus likely the result of stomatal closure as the result of downregulating mechanisms due to leaf-level water stress. As reported in previous studies, there could be an imbalance between water demand and supply in leaves if leaf water storage is depleted faster than it is being recharged through stems and roots (Brodribb and Holbrook, 2007; Matheny et al., 2014; Sperry et al., 2002). This could result in an eventual depletion of trunk water storage reservoirs in oil palms, which could lead to down-regulations of stomatal conductance (Damour et al., 2010; McCulloh et al., 2012; Zhang et al., 2014) and thus explain the observed decreases in T and GPP early in the day. The depleted storage would subsequently be refilled during nighttime hours (Sperling et al., 2014; Yang et al., 2012), when there is no leaf T. Such trunk water storage mechanisms have been described for some tree species (Goldstein et al., 1998; Waring et al., 1979; Waring and Running, 1978) and substantial trunk water storage capacities have been reported for several palm species (Holbrook and Sinclair, 1992; Madurapperuma et al., 2009; Sperling et al., 2014). Our data provide indications that internal water storage could be contributing to water fluxes in mature oil palm plantations, but more research is needed to confirm or reject this hypothesis. The high ET rates measured in the evening in the mature plantation, when T is already relatively low (Fig. 5b), suggest that evaporation is high particularly during the later hours of the day. Water storage capacity of stem surfaces of tropical trees can exceed 50% of the total interception storage (Herwitz, 1985), particularly when there is some epiphyte load on the stems. The trunk surface storage capacity of oil palms could potentially be even larger due to axils that remain on the trunk for years following the constant manual pruning of leaves; they constitute water reservoirs from which significant amounts of water can evaporate. Simulations per-

81

formed with the CLM-Palm model (Fan et al., 2015) showed that increasing the water interception efficiency and storage capacity of the canopy, and particularly that of trunk surfaces, resulted in a much better agreement between measured and modeled H, LE, ET and T fluxes (Fig. 6a and b; Appendix Table A3). In our simulations, the water interception efficiency (fpimx) was thus similar to that in the first versions of the CLM (e.g. CLM3; Oleson et al., 2004), before being decreased to match global patterns of ET partitioning in follow-up versions of the CLM (Lawrence et al., 2007). Additionally, the simulated fluxes were significantly improved by recognizing the potentially larger water storage capacity of trunk surfaces than leaf surfaces when incorporating distinct water storage parameters (dewmx1, dewmx2) for each of them. This confirms our hypothesis that increasing the water storage capacity of oil palm leaf axils along the trunk allows for a longer-lasting evaporation after rain events, which further improved the agreement of modeled afternoon energy fluxes with measured data (Fig. 6b) compared to using one single dewmx (Appendix Table A3). The independent treatment of trunk and leaf surfaces also avoids overfitting a single water storage parameter for leaves (dewmx1) to the observed fluxes. The series of model experiments highlight the importance of recognizing different precipitation interception efficiencies and storage capacities between global averages of various PFTs and those specific to a certain PFT, particularly for tropical ecosystems. Precipitation interception has been reported to be significantly higher for rainforests than in other forest types (Dykes, 1997; Lockwood and Sellers, 1982). The canopy of mature oil palms has an interception efficiency and storage capacity similar to that of forests according to measurements by Dufrêne et al. (1992). The revised parameterization with fpimx = 1, dewmx1 = 0.4 mm and dewmx2 = 6 mm (Appendix Table A3) gave the best estimate of the annual rainfall to interception ratio (18.5%); it is comparable to values previously reported for oil palm (21%; Nelson et al., 2014). This suggests that parameters controlling water interception in the CLM (i.e. fpimx and dewmx) should be adapted to species-specific canopy traits such as the substantial evaporation from leaf axil water reservoirs on oil palm trunks. Our new canopy hydrological parameterization and benchmarking against observation show that using water interception efficiencies and capacities that are specific to certain PFTs instead of relying on common default parameterization for all species can significantly improve the overall accuracy of model predictions.

5. Conclusions Our study provides first ecosystem-scale measurements of water and energy fluxes in oil palm plantations and evaluates the effects of plantation age. In oil palm plantations, most of the available energy is used for evapotranspiration rather than for sensible heat fluxes, particularly in the studied mature plantation (12-years old). In the young plantation (1-year old), sensible heat fluxes were higher. These results point to potentially substantial consequences regarding energy and water budgets when taking a look at the global picture and particularly at the continuing rapid conversion of different tropical ecosystems to oil palm plantations. Oil palm plantations at young stages of development have the potential to lead to surface heating due to decreased evaporative cooling and increased sensible heat fluxes. These warming effects weaken when the oil palm plantation becomes older due to increased latent heat fluxes. Nevertheless, the 25 year rotation cycle of oil palms creates landscapes where there is always a substantial part of young plantations, and thus warmer areas. Consequently, the large-scale conversion to agricultural monocultures such as oil palm plantations induces potentially substantial surface heating additionally to climate change at different spatial scales.

82

A. Meijide et al. / Agricultural and Forest Meteorology 239 (2017) 71–85

Trunk surfaces of mature palms provide external water reservoirs, namely in the axils of pruned leaf petioles. They intercept and store substantial amounts of precipitated water, which significantly contribute to evaporative water fluxes. Accounting for this oil palm specific interception trait in the CLM-Palm model by increasing canopy water interception parameters resulted in an improved fit between measured and modeled fluxes including the non-stable diurnal partitioning of energy fluxes. To improve the predictions of land-use driven climatic changes, there is a need for ecosystem-specific parametrization of the currently applied models based on data obtained from field measurements. We provide first estimates of albedos, WUEs, annual T and ET rates and of cropping coefficients for deriving ET based on the Penman-Monteith-FAO equation in young and mature plantations. Some of the values presented and processes described in our study may be valuable for the parametrization of such models, which can contribute to the further understanding of regional and global climatic effects of oil palm expansion.

Acknowledgements This study was financed by the Deutsche Forschungsgemeinschaft (DFG) in the framework of the collaborative GermanIndonesian research project CRC990 (subprojects A02 and A03). Yuanchao Fan was funded by an Erasmus Mundus FONASO Doctorate fellowship from the European Commission. The authors would like to thank Heri Junedi from the University of Jambi (UNJA) for his collaboration with the project and Dodo Gunawan from the Indonesian Meteorological Service (BMKG) for providing long-term climatic data. We would also like to thank Thomas Guillaume from the University of Göttingen, for information on soil properties in the two plantations. Special thanks to our field assistants in Indonesia and Edgar Tunsch, Dietmar Fellert, and Malte Puhan for technical assistance. We also thank PTPN VI and the owner of the plantation at Pompa Air for allowing us to conduct our research at their plantation. Appendix A.

Table A1 Measurement periods for different techniques and locations. Eddy covariance

Sap flux

Meteorology

Vcmax, Jmax, gmax, gs

1-year old (Pompa Air) 12-year old (PTPN VI)

01/07/2013−18/02/2014 30/05/2014−03/02/2015

01/07/2013–18/02/2014 03/03/2014–05/05/2015

12/2015a 12/2015

Reference meteorological station − Reki

—

18/10/2013−12/01/2014 20/05/2014 3–10/06/2014 19–27/07/2014 21/10/2014 30/10/2014−20/12/2014 —

a

11/12/2013–10/12/2014

Measurements were performed in a 1-year old plantation located at <2 km from Pompa Air and with similar management practices.

Table A2 Number of 30 min data points and percentages of available data in relation to the measured data (in brackets) after different filtering steps during eddy covariance analysis.

Measurement period Measured data Quality filtering and integral turbulence characteristics Ustar filtering Footprint analysis Filtering for outliers/Final

1-year old (Pompa Air)

12-year old (PTPN VI)

11134 7063 6036 (85.5%) 4357 (61.7%) 4144 (58.7%) 4128 (58.4%)

11977 8940 7546 (84.4%) 6220 (69.6%) 6202 (69.4%) 6149 (68.8%)

Table A3 Root mean squared error (RMSE) of simulated water and energy fluxes compared to observed fluxes, averaged from all available half-hourly values from June to December 2014. Simulations were conducted using CLM-Palm by a series of canopy interception experiments as described in Fan (2016). Experiment

LE (W m−2 )

H (W m−2 )

ET (mm h−1 )

T (mm h−1 )

fpimx = 0.25; dewmx = 0.1 fpimx = 0.5; dewmx = 0.2 fpimx = 0.5; dewmx = 0.4 fpimx = 1; dewmx = 0.2 fpimx = 1; dewmx = 0.4 fpimx = 1; dewmx 1 = 0.2; dewmx 2 = 2; fpimx = 1; dewmx 1 = 0.4; dewmx 2 = 6;

39.7 38.2 35.6 37.9 34.9 34.2

20.3 19.5 18.4 19.3 18.1 17.0

0.064 0.062 0.058 0.061 0.057 0.056

0.056 0.055 0.051 0.055 0.051 0.048

29.5

14.6

0.049

0.043

A. Meijide et al. / Agricultural and Forest Meteorology 239 (2017) 71–85

References Ali, A.A., Xu, C., Rogers, A., McDowell, N.G., Medlyn, B.E., Fisher, R.A., Wullschleger, S.D., Reich, P.B., Vrugt, J.A., Bauerle, W.L., Santiago, L.S., Wilson, C.J., 2015. Global scale environmental control of plant photosynthetic capacity. Ecol. Appl. 25, http://dx.doi.org/10.1890/14-2111.1, 150511125956002. Allen, R.G, Pereira, L.S., Raes, D. and Smith,M., 1998: Crop evapotranspiration Guidelines for computing crop water requirements - FAO Irrigation and drainage paper 56, FAO, Rome 300, no. 9 (1998): D05109. Alkama, R., Cescatti, A., 2016. Biophysical climate impacts of recent changes in global forest cover. Science 351 (6273), 600–603. Allen, K., Corre, M.D., Tjoa, A., Veldkamp, E., 2015. Soil nitrogen-Cycling responses to conversion of lowland forests to oil palm and rubber plantations in sumatra, Indonesia. PLoS One 10, e0133325, http://dx.doi.org/10.1371/journal.pone. 0133325. Aragao, L.E.O.C., 2012. Environmental science: the rainforest’s water pump. Nature 489, 217–218. Aubinet, M., Grelle, A., Ibrom, A., Rannik, Ü., Moncrieff, J., Foken, T., Kowalski, A.S., Martin, P.H., Berbigier, P., Bernhofer, C., Clement, R., Elbers, J., Granier, A., Grünwald, T., Morgenstern, K., Pilegaard, K., Rebmann, C., Snijders, W., Valentini, R., Vesala, T., 1999. Estimates of the annual net carbon and water exchange of forests: the EUROFLUX methodology. Adv. Ecol. Res. 30, 113–175, http://dx.doi.org/10.1016/S0065-2504(08)60018-5. Awal, M.A., Wan Ishak, W.I., Bockari-Gevao, S.M., 2010. Determination of leaf area index for oil palm plantation using hemispherical photography technique. Pertanika J. Sci. Technol. 18, 23–32. Baldocchi, D., 1997. Measuring and modelling carbon dioxide and water vapour exchange over a temperate broad-leaved forest during the 1995 summer drought. Plant. Cell Environ. 20, 1108–1122, http://dx.doi.org/10.1046/j.13653040.1997.d01-147.x. Barnes, A.D., Jochum, M., Mumme, S., Haneda, N.F., Farajallah, A., Widarto, T.H., Brose, U., 2014. Consequences of tropical land use for multitrophic biodiversity and ecosystem functioning. Nat. Commun. 5, 5351, http://dx.doi.org/10.1038/ ncomms6351. Bastable, H.G., Shuttleworth, W.J., Dallarosa, R.L.G., Fisch, G., Nobre, C.A., 1993. Observations of climate, albedo, and surface radiation over cleared and undisturbed amazonian forest. Int. J. Climatol. 13, 783–796, http://dx.doi.org/ 10.1002/joc.3370130706. Berbet, M.L.C., Costa, M.H., 2003. Climate change after tropical deforestation: seasonal variability of surface albedo and its effects on precipitation change. J. Clim. 16, 2099–2104, http://dx.doi.org/10.1175/15200442(2003)016<2099:CCATDS>2.0.CO;2. Betts, R.A., 2000. Offset of the potential carbon sink from boreal forestation by decreases in surface albedo. Nature 408, 187–190, http://dx.doi.org/10.1038/ 35041545. Bounoua, L., DeFries, R., Collatz, G.J., Sellers, P., Khan, H., 2002. Effects of land cover conversion on surface climate. Clim. Change 52, 29–64, http://dx.doi.org/10. 1023/A:1013051420309. Brodribb, T.J., Holbrook, N.M., 2007. Forced depression of leaf hydraulic conductance in situ: effects on the leaf gas exchange of forest trees. Funct. Ecol. 21, 705–712, http://dx.doi.org/10.1111/j.1365-2435.2007.01271.x. Carlson, K.M., Curran, L.M., Ratnasari, D., Pittman, A.M., Soares-Filho, B.S., Asner, G.P., Trigg, S.N., Gaveau, D., Lawrence, D., 2012. Committed carbon emissions, deforestation, and community land conversion from oil palm plantation expansion in West Kalimantan. Indonesia. Proc. Natl. Acad. Sci. U. S. A. 109, 7559–7564, http://dx.doi.org/10.1073/pnas.1200452109. Carr, M.K.V., 2011. The water relations and irrigation requirements of oil palm (Elaeis guineensis): a review. Exp. Agricul. 47, 629–652, http://dx.doi.org/10. 1017/S0014479711000494 793. Chapin, F.S., Woodwell, G.M., Randerson, J.T., Rastetter, E.B., Lovett, G.M., Baldocchi, D.D., Clark, D.A., Harmon, M.E., Schimel, D.S., Valentini, R., Wirth, C., Aber, J.D., Cole, J.J., Goulden, M.L., Harden, J.W., Heimann, M., Howarth, R.W., Matson, P.A., McGuire, A.D., Melillo, J.M., Mooney, H.A., Neff, J.C., Houghton, R.A., Pace, M.L., Ryan, M.G., Running, S.W., Sala, O.E., Schlesinger, W.H., Schulze, E.D., 2006. Reconciling carbon-cycle concepts, terminology, and methods. Ecosystems 9, 1041–1050, http://dx.doi.org/10.1007/s10021-005-0105-7. Chaves, M.M., 2004. Mechanisms underlying plant resilience to water deficits: prospects for water-saving agriculture. J. Exp. Bot. 55, 2365–2384, http://dx. doi.org/10.1093/jxb/erh269. Collatz, G.J., Ball, J.T., Grivet, C., Berry, J.A., 1991. Physiological and environmental regulation of stomatal conductance, photosynthesis and transpiration: a model that includes a laminar boundary layer. Agric. For. Meteorol. 54, 107–136, http://dx.doi.org/10.1016/0168-1923(91)90002-8. Comte, I., Colin, F., Whalen, J.K., Grünberger, O., Caliman, J.P., 2012. Agricultural practices in oil palm plantations and their impact on hydrological changes, nutrient fluxes and water quality in Indonesia. a review. Adv. Agron., http://dx. doi.org/10.1016/B978-0-12-394277-7.00003-8. Culf, A.D., Fisch, G., Hodnett, M.G., 1995. The albedo of amazonian forest and ranch land. J. Clim. 8, 1544–1554, http://dx.doi.org/10.1175/15200442(1995)008<1544:TAOAFA>2.0.CO;2. Damour, G., Simonneau, T., Cochard, H., Urban, L., 2010. An overview of models of stomatal conductance at the leaf level. Plant. Cell Environ. 33, 1419–1438, http://dx.doi.org/10.1111/j.1365-3040.2010.02181.x. De Kauwe, M.G., Medlyn, B.E., Zaehle, S., Walker, A.P., Dietze, M.C., Hickler, T., Jain, A.K., Luo, Y., Parton, W.J., Prentice, I.C., Smith, B., Thornton, P.E., Wang, S., Wang, Y.-P., Wårlind, D., Weng, E., Crous, K.Y., Ellsworth, D.S., Hanson, P.J., Seok

83

Kim, H.-, Warren, J.M., Oren, R., Norby, R.J., 2013. Forest water use and water use efficiency at elevated CO2: a model-data intercomparison at two contrasting temperate forest FACE sites. Glob. Change Biol. 19, 1759–1779, http://dx.doi.org/10.1111/gcb.12164. Defries, R.S., Bounoua, L., Collatz, G.J., 2002. Human modification of the landscape and surface climate in the next fifty years. Glob. Change Biol. 8, 438–458, http://dx.doi.org/10.1046/j.1365-2486.2002.00483.x. Dislich, C., Keyel, A.C., Salecker, J., Kisel, Y., Meyer, K.M., Auliya, M., Barnes, A.D., Corre, M.D., Darras, K., Faust, H., Hess, B., Klasen, S., Knohl, A., Kreft, H., Meijide, A., Nurdiansyah, F., Otten, F., Pe’er, G., Steinebach, S., Tarigan, S., Tölle, M.H., Tscharntke, T., Wiegand, K., 2016. A review of the ecosystem functions in oil palm plantations, using forests as a reference system. Biol. Rev. 49, http://dx. doi.org/10.1111/brv.12295. Domingo, F., Villagarcía, L., Brenner, A.J., Puigdefábregas, J., 2000. Measuring and modelling the radiation balance of a heterogeneous shrubland. Plant. Cell Environ. 23, 27–38, http://dx.doi.org/10.1046/j.1365-3040.2000.00532.x. Drescher, J., Rembold, K., Allen, K., Beckscha, P., Buchori, D., Clough, Y., Faust, H., Fauzi, A.M., Gunawan, D., Hertel, D., Irawan, B., Jaya, I.N.S., Klarner, B., Kleinn, C., Knohl, A., Kotowska, M.M., Krashevska, V., Krishna, V., Leuschner, C., Lorenz, W., Meijide, A., Melati, D., Steinebach, S., Tjoa, A., Tscharntke, T., Wick, B., Wiegand, K., Kreft, H., Scheu, S., 2016. Ecological and socio-economic functions across tropical land use systems after rainforest conversion., http://dx.doi.org/ 10.1098/rstb.2015.0275. Duursma, R.A., 2015. Plantecophys − An R Package for Analysing and Modelling Leaf Gas Exchange Data. PLoS One 10, e0143346, http://dx.doi.org/10.1371/ journal.pone.0143346. Dykes, A., 1997. Rainfall interception from a lowland tropical rainforest in Brunei. J. Hydrol. 200, 260–279. Ehleringer, J.R., 1993. Carbon and water relations in desert plants: an isotopic perspective. In: Ehleringer, J.R., Hall, A.E., Farquhar, G.D. (Eds.), Stable Isotopes and Plant Carbon–Water Relations. Academic Press, San Diego, pp. 155–172. Escudero, A., Mediavilla, S., 2003. Decline in photosynthetic nitrogen use efficiency with leaf age and nitrogen resorption as determinants of leaf life span. J. Ecol. 91, 880–889, http://dx.doi.org/10.1046/j.1365-2745.2003.00818.x. Eugster, W., Plüss, P., 2010. A fault-tolerant eddy covariance system for measuring CH4 fluxes. Agric. For. Meteorol. 150, 841–851, http://dx.doi.org/10.1016/j. agrformet.2009.12.008. FAO, 2012. Food and Agriculture Organization of the United Nations FAOSTAT, available at: http://faostat.fao.org. Falge, E., Baldocchi, D., Olson, R., Anthoni, P., Aubinet, M., Bernhofer, C., Burba, G., Ceulemans, R., Clement, R., Dolman, H., Granier, A., Gross, P., Grünwald, T., Hollinger, D., Jensen, N.-O., Katul, G., Keronen, P., Kowalski, A., Lai, C.T., Law, B.E., Meyers, T., Moncrieff, J., Moors, E., Munger, J.W., Pilegaard, K., Rannik, Ü., Rebmann, C., Suyker, A., Tenhunen, J., Tu, K., Verma, S., Vesala, T., Wilson, K., Wofsy, S., 2001. Gap filling strategies for defensible annual sums of net ecosystem exchange. Agric. For. Meteorol. 107, 43–69, http://dx.doi.org/10. 1016/S0168-1923(00)00225-2. Fan, Y., 171 pp.; available at: 2016. Modeling Oil Palm Monoculture and Its Associated Impacts on Land-atmosphere Carbon, Water and Energy Fluxes in Indonesia. Ph.D. Thesis. University of Göttingen, Germany http://hdl.handle. net/11858/00-1735-0000-0028-876B-4. Fan, Y., Roupsard, O., Bernoux, M., Le Maire, G., Panferov, O., Kotowska, M.M., Knohl, A., 2015. A sub-canopy structure for simulating oil palm in the Community Land Model (CLM-Palm): Phenology, allocation and yield. Geosci. Model Dev. 8, 3785–3800, http://dx.doi.org/10.5194/gmd-8-3785-2015. Fowler, D., Nemitz, E., Misztal, P., Di Marco, C., Skiba, U., Ryder, J., Helfter, C., Cape, J.N., Owen, S., Dorsey, J., Gallagher, M.W., Coyle, M., Phillips, G., Davison, B., Langford, B., MacKenzie, R., Muller, J., Siong, J., Dari-Salisburgo, C., Di Carlo, P., Aruffo, E., Giammaria, F., Pyle, J.A., Hewitt, C.N., 2011. Effects of land use on surface-atmosphere exchanges of trace gases and energy in Borneo: comparing fluxes over oil palm plantations and a rainforest. Philos. Trans. R. Soc. B Biol. Sci. 366, 3196–3209, http://dx.doi.org/10.1098/rstb.2011.0055. Gash, J.C., Shuttleworth, W.J., 1991. Tropical deforestation: albedo and the surface-energy balance. Clim. Change 19, 123–133, http://dx.doi.org/10.1007/ BF00142219. Goldstein, G., Andrade, J.L., Meinzer, F.C., Holbrook, N.M., Cavelier, J., Jackson, P., Celis, A., 1998. Stem water storage and diurnal patterns of water use in tropical forest canopy trees. Plant. Cell Environ. 21, 397–406, http://dx.doi.org/10. 1046/j.1365-3040.1998.00273.x. Goulden, M.L., Munger, J.W., Fan, S.-M., Daube, B.C., Wofsy, S.C., 1996. Measurements of carbon sequestration by long-term eddy covariance: methods and a critical evaluation of accuracy. Glob. Change Biol. 2, 169–182, http://dx.doi.org/10.1111/j.1365-2486.1996.tb00070.x. Grace, J., Grace, J., Lloyd, J., Lloyd, J., McIntyre, J., McIntyre, J., Miranda a. Miranda a. Meir, P., Meir, P., Miranda, H., Miranda, H., Moncrieff, J., Moncrieff, J., Massheder, J., Massheder, J., Wright, I., Wright, I., Gash, J., Gash, J., 1995. Fluxes of carbon dioxide and water vapour over an undisturbed tropical forest in south-west Amazonia. Glob. Change Biol. 1, 1–12. Granier, A., Biron, P., Lemoine, D., 2000. Water balance, transpiration and canopy conductance in two beech stands. Agric. For. Meteorol. 100, 291–308, http:// dx.doi.org/10.1016/S0168-1923(99)00151-3. Granier, A., 1985. Une nouvelle méthode pour la mesure du flux de sève brute dans le tronc des arbres. Ann. For. Sci. 42, 193–200. Guillaume, T., Muhammad, D., Kuzyakov, Y., 2015. Losses of soil carbon by converting tropical forest to plantations: erosion and decomposition estimated by ␦13C. Glob. Chang. Biol., http://dx.doi.org/10.1111/gcb.12907.

84

A. Meijide et al. / Agricultural and Forest Meteorology 239 (2017) 71–85

Han, Q., Kawasaki, T., Nakano, T., Chiba, Y., 2008. Leaf-age effects on seasonal variability in photosynthetic parameters and its relationships with leaf mass per area and leaf nitrogen concentration within a Pinus densiflora crown. Tree Physiol. 28, 551–558, http://dx.doi.org/10.1093/treephys/28.4.551. Hardwick, S.R., Toumi, R., Pfeifer, M., Turner, E.C., Nilus, R., Ewers, R.M., 2015. The relationship between leaf area index and microclimate in tropical forest and oil palm plantation: forest disturbance drives changes in microclimate. Agric. For. Meteorol. 201, 187–195, http://dx.doi.org/10.1016/j.agrformet.2014.11.010. Henson, I.E., Harun, M.H., 2005. The influence of climatic conditions on gas and energy exchanges above a young oil palm stand in North Kedah. Oil Plam Bull. 17, 73–91. Herwitz, S.R., 1985. Interception storage capacities of tropical rainforest canopy trees. J. Hydrol. 77, 237–252. Heuzé, V., Tran, G., Sauvant, D., Lebas, F., 2015. Oil palm fronds and oil palm crop residues. Feedipedia, a programme by INRA, CIRAD, AFZ and FAO. http://www. feedipedia.org/node/6916 Last updated on June 25, 2015, 11:18. Holbrook, N.M., Sinclair, T.R., 1992. Water balance in the arborescent palm, Sabal palmetto. II. Transpiration and stem water storage. Plant. Cell Environ. 15, 401–409, http://dx.doi.org/10.1111/j.1365-3040.1992. tb00990.x. Hollinger, D.Y., Ollinger, S., Richardson V, A.D., Meyers, T.P., Dail, D.B., Martin, M.E., Scott, N.A., Arkebauer, T.J., Baldocchi, D.D., Clark, K.L., Curtis, P.S., Davis, K.J., Desai, A.R., Dragoni, D., Goulden, M.L., Gu, L., Katul, G.G., Pallardy, S.G., Paw U, K.T., Schmid, H.P., Stoy, P.C., Suker, A.E., Verma, S.B., 2010. Albedo estimates for land surface models and support for a new paradigm based on foliage nitrogen concentration. Glob. Change Biol. 16, 696–710, http://dx.doi.org/10.1111/j. 1365-2486.2009.02028.x. Humphreys, E.R., Black, T.A., Ethier, G.J., Drewitt, G.B., Spitlehouse, D.L., Jork, E.M., Nesic, Z., Livingston, N.J., 2003. Annual and seasonal variability of sensible and latent heat fluxes abpve a coastal Douglas-fir forest, British Columbia, Canada. Agric. For. Meteorol. 115, 109–125. Islam, M., Dahlan, I., Rajion, M.A., Jelan, Z.A., 2000. Productivity and nutritive values of different fractions of oil palm (Elaeis guineensis) frond. Asian Aust. J. Anim. Sci. 13, 1113–1120, http://dx.doi.org/10.5713/ajas.2000.1113. Jones, H.G., Higgs, K.H., 1989. Empirical models of the conductance of leaves in apple orchards. Plant. Cell Environ. 12, 301–308, http://dx.doi.org/10.1111/j. 1365-3040.1989. tb01944.x. Kattge, J., Knorr, W., Raddatz, T., Wirth, C., 2009. Quantifying photosynthetic capacity and its relationship to leaf nitrogen content for global-scale terrestrial biosphere models. Glob. Change Biol. 15, 976–991, http://dx.doi.org/10.1111/j. 1365-2486.2008.01744.x. Keenan, T.F., Hollinger, D.Y., Bohrer, G., Dragoni, D., Munger, J.W., Schmid, H.P., Richardson, A.D., 2013. Increase in forest water-use efficiency as atmospheric carbon dioxide concentrations rise. Nature 499, 324–327. Kljun, N., Calanca, P., Rotach, M.W., Schmid, H.P., 2004. A simple parameterisation for flux footprint predictions. Boundary-Layer Meteorol. 112, 503–523, http:// dx.doi.org/10.1023/B:BOUN.0000030653.71031.96. Knohl, A., Buchmann, N., 2005. Partitioning the net CO2 flux of a deciduous forest into respiration and assimilation using stable carbon isotopes. Glob. Biogeochem. Cycles 19, 1–14, http://dx.doi.org/10.1029/2004GB002301. Kositsup, B., Kasemsap, P., Thanisawanyangkura, S., Chairungsee, N., Satakhun, D., Teerawatanasuk, K., Ameglio, T., Thaler, P., 2010. Effect of leaf age and position on light-saturated CO2 assimilation rate, photosynthetic capacity, and stomatal conductance in rubber trees. Photosynthetica 48, 67–78, http://dx. doi.org/10.1007/s11099-010-0010-y. Kotowska, M.M., Leuschner, C., Triadiati, T., Meriem, S., Hertel, D., 2015. Quantifying above- and belowground biomass carbon loss with forest conversion in tropical lowlands of Sumatra (Indonesia). Glob. Change Biol., http://dx.doi.org/10.1111/gcb.12979. Kramm, G., Dlugi, R., Mölders, N., 2002. Sublayer-Stanton numbers of heat and matter for aerodynamically smooth surfaces: basic considerations and evaluation. Meteorol. Atmos. Phys. 79, 173–194, http://dx.doi.org/10.1007/ s007030200002. Kumagai, T., Saitoh, T.M., Sato, Y., Takahashi, H., Manfroi, O.J., Morooka, T., Kuraji, K., Suzuki, M., Yasunari, T., Komatsu, H., 2005. Annual water balance and seasonality of evapotranspiration in a Bornean tropical rainforest. Agric. For. Meteorol. 128, 81–92, http://dx.doi.org/10.1016/j.agrformet.2004.08.006. Laurila, T., Tuovinen, J.-P., Lohila, A., Hatakka, J., Aurela, M., Thum, T., Pihlatie, M., Rinne, J., Vesala, T., 2005. Measuring methane emissions from a landfill using a cost-effective micrometeorological method. Geophys. Res. Lett. 32, http://dx. doi.org/10.1029/2005gl023462. Lawrence, D.M., Thornton, P.E., Oleson, K.W., Bonan, G.B., 2007. The partitioning of evapotranspiration into transpiration, soil evaporation, and canopy evaporation in a GCM: impacts on land-atmosphere interaction. J. Hydrometeorol. 8, 862–880, http://dx.doi.org/10.1175/jhm596.1. Li, S.-G., Eugster, W., Asanuma, J., Kotani, A., Davaa, G., Oyunbaatar, D., Sugita, M., 2006. Energy partitioning and its biophysical controls above a grazing steppe in central Mongolia. Agric. For. Meteorol. 137, 89–106, http://dx.doi.org/10. 1016/j.agrformet.2006.03.010. Lockwood, J.G., Sellers, P.J., 1982. Comparisons of interception loss from tropical and temperate vegetation canopies. J. Appl. Meteorol. 21, 1405–1412, http:// dx.doi.org/10.1175/1520-0450(1982)021<1405:COILFT>2.0.CO;2. Luis, Z.G., Bezerra, K.M.G., Scherwinski-Pereira, J.E., 2010. Adaptability and leaf anatomical features in oil palm seedlings produced by embryo rescue and pre-germinated seeds. Braz. J. Plant Physiol. 22, 209–2015, 00.0000/S00000-000-0000-0.

Luskin, M.S., Potts, M.D., 2011. Microclimate and habitat heterogeneity through the oil palm lifecycle. Basic Appl. Ecol. 12, 540–551, http://dx.doi.org/10.1016/j. baae.2011.06.004. Madurapperuma, W.S., Bleby, T.M., Burgess, S.S.O., 2009. Evaluation of sap flow methods to determine water use by cultivated palms. Environ. Exp. Bot. 66, 372–380, http://dx.doi.org/10.1016/j.envexpbot.2009.04.002. Margono, B.A., Potapov, P., Turubanova V, S., Stolle, F., Hansen, M.C., 2014. Primary forest cover loss in Indonesia over 2000–2012. Nat. Clim. Change 4, 1–6, http:// dx.doi.org/10.1038/NCLIMATE2277. Matheny, A.M., Bohrer, G., Stoy, P.C., Baker, I.T., Black, A.T., Desai, A.R., Dietze, M.C., Gough, C.M., Ivanov, V.Y., Jassal, R.S., Novick, K.a., Schäfer, K.V.R., Verbeeck, H., 2014. Characterizing the diurnal patterns of errors in the prediction of evapotranspiration by several land-surface models: an NACP analysis. J. Geophys. Res. Biogeosci., http://dx.doi.org/10.1002/2014jg002623. McCulloh, K.A., Johnson, D.M., Meinzer, F.C., Voelker, S.L., Lachenbruch, B., Domec, J.C., 2012. Hydraulic architecture of two species differing in wood density: opposing strategies in co-occurring tropical pioneer trees. Plant Cell Environ. 35, 116–125, http://dx.doi.org/10.1111/j.1365-3040.2011.02421.x. Merten, J., Röll, A., Guillaume, T., Meijide, A., Tarigan, S., Agusta, H., Dislich, C., Dittrich, C., Faust, H., Gunawan, D., Hein, J., Hendrayanto, H., Knohl, A., Kuzyakov, Y., Wiegand, K., Hölscher, D., 2016. Water scarcity and oil palm expansion: social views and environmental processes. Ecol. Soc. 21 (2), 5, http://dx.doi.org/10.5751/es-08214-210205. Moncrieff, J.B., Massheder, J.M., de Bruin, H., Elbers, J., Friborg, T., Heusinkveld, B., Kabat, P., Scott, S., Soegaard, H., Verhoef, A., 1997. A system to measure surface fluxes of momentum, sensible heat, water vapour and carbon dioxide. J. Hydrol. 188–189, 589–611, http://dx.doi.org/10.1016/S0022-1694(96)03194-0. Moncrieff, J., Clement, R., Finnigan, J., Meyers, T., 2005. Averaging, detrending, and filtering of eddy covariance time series. In: Lee, X., Massman, W., Law, B. (Eds.), Handbook of Micrometeorology SE − 2. Atmospheric and Oceanographic Sciences Library, Springer, Netherlands, pp. 7–31, http://dx.doi.org/10.1007/14020-2265-4 2. Nelson, P., Webb, M., Banabas, M., Nake, S., Goodrick, I., Gordon, J., O’Grady, D., Dubos, B., 2014. Methods to account for tree-scale variability in soil-and plant-related parameters in oil palm plantations. Plant Soil 374, 459–471. Niinemets, Ü., 2002. Stomatal conductance alone does not explain the decline in foliar photosynthetic rates with increasing tree age and size in Picea abies and Pinus sylvestris. Tree Physiol. 22, 515–535, http://dx.doi.org/10.1093/ treephys/22.8.515. Niu, F., Röll, A., Hardanto, A., Meijide, A., Kohler, M., Holscher, Hendrayanto, 2015. Oil palm water use: calibration of a sap flux method and a field measurement scheme. Tree Physiol. 1 -11, http://dx.doi.org/10.1093/treephys/tpv013. OECD-FAO, Agricultural Outlook, 2012. Chp. 5 Oilseeds and oilseed products. Oishi, A.C., Oren, R., Stoy, P.C., 2008. Estimating components of forest evapotranspiration: a footprint approach for scaling sap flux measurements. Agric. For. Meteorol. 148, 1719–1732, http://dx.doi.org/10.1016/j.agrformet. 2008.06.013. Oleson, K.W., Dai, Y., Bonan, G., Bosilovich, M., Dickinson, R., Dirmeyer, P., Hoffman, F., Houser, P., Levis, S., Niu, G.-Y., 2004. Technical description of the community land model (CLM). Tech. Note NCAR/TN-461+ STR. Oleson, K., Lawrence, D., Bonan, G., Drewniak, B., Huang, M., Koven, C., Levis, S., Li, F., Riley, W., Subin, Z., Swenson, S., Thornton, P., Bozbiyik, A., Fisher, R., Heald, C., Kluzek, E., Lamarque, J.-F., Lawrence, P., Leung, L., Lipscomb, W., Muszala, S., Ricciuto, D., Sacks,W., Sun, Y., Tang, J., and Yang, Z.-L.: Technical description of version 4.5 of the Community Land Model (CLM), National Center for Atmospheric Research,Boulder, Colorado, USA, 420 pp., doi:10.5065/D6RR1W7M, 2013. Osmond, C.B., Winter, K., Ziegler, Z., 1982. Functional significance of different pathways of CO2 fixation in photosynthesis. In: Person, A., Zimmerman, H. (Eds.), Encyclopedia of Plant Physiology. Springer-Verlag, New York, Vol. 12B. Pielke, R., a, S., Avissar, R., Raupach, M.R., Dolman, a J., Zeng, X., Denning, S., 1998. Interactions between the atmosphere and terestrial ecosystems: influence on weather and climate. Glob. Change Biol. 4, 461–475. Ponton, S., Flanagan, L.B., Alstad, K.P., Johnson, B.G., Morgenstern, K., Kljun, N., Black, T.A., Barr, A.G., 2006. Comparison of ecosystem water-use efficiency among Douglas-fir forest, aspen forest and grassland using eddy covariance and carbon isotope techniques. Glob. Change Biol. 12, 294–310, http://dx.doi. org/10.1111/j.1365-2486.2005.01103.x. R Core Team, 2014. R: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing, Vienna, Austria, URL http://www.Rproject.org/. Röll, A., Niu, F., Meijide, A., Hardanto, A., Hendrayanto, Knohl, A., Hölscher, D., 2015. Transpiration in an oil palm landscape: effects of palm age. Biogeosciences 12, 5619–5633, http://dx.doi.org/10.5194/bg-12-5619-2015. Reichstein, M., Falge, E., Baldocchi, D., Papale, D., Aubinet, M., Berbigier, P., Bernhofer, C., Buchmann, N., Gilmanov, T., Granier, A., Grunwald, T., Havrankova, K., Ilvesniemi, H., Janous, D., Knohl, A., Laurila, T., Lohila, A., Loustau, D., Matteucci, G., Meyers, T., Miglietta, F., Ourcival, J.-M., Pumpanen, J., Rambal, S., Rotenberg, E., Sanz, M., Tenhunen, J., Seufert, G., Vaccari, F., Vesala, T., Yakir, D., Valentini, R., 2005. On the separation of net ecosystem exchange into assimilation and ecosystem respiration: review and improved algorithm. Glob. Change Biol. 11, 1424–1439, http://dx.doi.org/10.1111/j.1365-2486. 2005.001002.x. Roupsard, O., Bonnefond, J.-M., Irvine, M., Berbigier, P., Nouvellon, Y., Dauzat, J., Taga, S., Hamel, O., Jourdan, C., Saint-André, L., Mialet-Serra, I., Labouisse, J.-P., Epron, D., Joffre, R., Braconnier, S., Rouzière, A., Navarro, M., Bouillet, J.-P., 2006.

A. Meijide et al. / Agricultural and Forest Meteorology 239 (2017) 71–85 Partitioning energy and evapo-transpiration above and below a tropical palm canopy. Agric. For. Meteorol. 139, 252–268, http://dx.doi.org/10.1016/j. agrformet.2006.07.006. Running, S.W., Baldocchi, D.D., Turner, D.P., Gower, S.T., Bakwin, P.S., Hibbard, K.A., 1999. A global terrestrial monitoring network integrating tower fluxes, flask sampling, ecosystem modeling and EOS satellite data. Remote Sens. Environ. 70, 108–127, http://dx.doi.org/10.1016/S0034-4257(99)00061-9. Schlesinger, W.H., Jasechko, S., 2014. Transpiration in the global water cycle. Agric. For. Meteorol. 189–190, 115–117, http://dx.doi.org/10.1016/j.agrformet.2014. 01.011. Schwaiger, H.P., Bird, D.N., 2010. Integration of albedo effects caused by land use change into the climate balance: should we still account in greenhouse gas units? For. Ecol. Manag. 260, 278–286, http://dx.doi.org/10.1016/j.foreco.2009. 12.002. Shuttleworth, W. J., Gash, J. H. C., Lloyd, C. R., Moore, C. J., Roberts, J., Filho, A. D. O. M., Fisch, G., De Paula Silva Filho, V., Ribeiro, M. D. N. G., Molion, L. C. B., De Abreu Sá, L. D., Nobre, J. C. A., Cabral, O. M. R., Patel, S. R. and Carvalho De Moraes, J., 1984. Eddy correlation measurements of energy partition for Amazonian forest. Q.J.R. Meteorol. Soc., 110: 1143–1162. 10.1002/qj.49711046622. Sperling, O., Shapira, O., Schwartz, A., Lazarovitch, N., 2014. Direct in vivo evidence of immense stem water exploitation in irrigated date palms. J. Exp. Bot., http:// dx.doi.org/10.1093/jxb/eru421. Sperry, J.S., Hacke, U.G., Oren, R., Comstock, J.P., 2002. Water deficits and hydraulic limits to leaf water supply. Plant. Cell Environ. 25, 251–263, http://dx.doi.org/ 10.1046/j.0016-8025.2001.00799.x. Stewart, J.B., 1977. Evaporation from the wet canopy of a pine forest. Water Resour. Res. 13, 915–921, http://dx.doi.org/10.1029/WR013i006p00915. Tan, Z.-H., Zhang, Y.-P., Deng, X.-B., Song, Q.-H., Liu, W.-J., Deng, Y., Tang, J.-W., Liao, Z.-Y., Zhao, J.-F., Song, L., Yang, L.-Y., 2015. Interannual and seasonal variability of water use efficiency in a tropical rainforest: results from a 9 year eddy flux time series. J. Geophys. Res. Atmos. 120, http://dx.doi.org/10.1002/ 2014JD022535 (2014JD022535). Tang, J., Bolstad, P.V., Ewers, B.E., Desai, A.R., Davis, K.J., Carey, E.V., 2006. Sap flux-upscaled canopy transpiration, stomatal conductance, and water use efficiency in an old growth forest in the Great Lakes region of the United States. J. Geophys. Res. Biogeosci. 111, 1–12, http://dx.doi.org/10.1029/2005JG000083. Tani, M., Nik, A.R., Yasuda, Y., Noguchi, S., Shamsuddin, S.A., Sahat, M.M., Takanashi, S., 2003. Long-term estimation of evapotranspiration from a tropical rain forest in Peninsular Malaysia, in Water Resources Systems - Water Availability and Global Change (Proceedings of an Symposium HS02a Held During IUGG 2003, the XXIII General Assembly of the International Union of Geodesy and Geophysics. Sapporo, Japan, IAHS Publ. no. 280, pp. 267.

85

Tenhunen, J.D., Valentini, R., Köstner, B., Zimmermann, R., Granier, A., 1998. Variation in forest gas exchange at landscape to continental scales. Ann. For. Sci. 55, 1–11. USDA, 2016. United States Department of Agriculture Indonesia: Oilseeds and Products Annual Report 2016, available at: http://gain.fas.usda.gov/ Recent%20GAIN%20Publications/Oilseeds%20and%20Products%20Annual Jakarta Indonesia 3-15-2016. pdf. van der Molen, M.K., Dolman, A.J., Waterloo, M.J., Bruijnzeel, L.A., 2006. Climate is affected more by maritime than by continental land use change: a multiple scale analysis. Glob. Planet Change 54, 128–149, http://dx.doi.org/10.1016/j. gloplacha.2006.05.005. Waring, R.H., Running, S.W., 1978. Sapwood water storage: its contribution to transpiration and effect upon water conductance through the stems of old-growth Douglas-fir. Plant. Cell Environ. 1, 131–140, http://dx.doi.org/10. 1111/j.1365-3040.1978.tb00754.x. Waring, R.H., Whitehead, D., Jarvis, P.G., 1979. The contribution of stored water to transpiration in Scots pine. Plant. Cell Environ. 2, 309–317, http://dx.doi.org/ 10.1111/j.1365-3040.1979.tb00085.x. Webb, E.K., Pearman, G.I., Leuning, R., 1980. Correction of flux measurements for density effects due to heat and water vapour transfer. Q. J. R. Meteorol. Soc. 106, 85–100, http://dx.doi.org/10.1002/qj.49710644707. Wilczak, J., Oncley, S., Stage, S., 2001. Sonic anemometer tilt correction algorithms. Boundary-Layer Meteorol. 99, 127–150, http://dx.doi.org/10.1023/ A:1018966204465. Wullschleger, S.D., 1993. Biochemical limitations to carbon assimilation in c 3 Plants—A retrospective analysis of the A/C i curves from 109 species. J. Exp. Bot. 44, 907–920, http://dx.doi.org/10.1093/jxb/44.5.907. Yang, S.-J., Zhang, Y.-J., Sun, M., Goldstein, G., Cao, K.-F., 2012. Recovery of diurnal depression of leaf hydraulic conductance in a subtropical woody bamboo species: embolism refilling by nocturnal root pressure. Tree Physiol., http://dx. doi.org/10.1093/treephys/tps028. Zhang, Q., Manzoni, S., Katul, G., Porporato, A., Yang, D., 2014. The hysteretic evapotranspiration—Vapor pressure deficit relation. J. Geophys. Res. Biogeosci. 119, 125–140, http://dx.doi.org/10.1002/2013JG002484.