Advertisement
Browse Subject Areas
?

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

  • Loading metrics

Rapid Leaf Deployment Strategies in a Deciduous Savanna

  • Edmund Carl February ,

    Contributed equally to this work with: Edmund Carl February, Steven Ian Higgins

    Edmund.February@uct.ac.za

    Affiliation Department of Biological Sciences, University of Cape Town, Private Bag, Rondebosch, 7701, South Africa

    ORCID http://orcid.org/0000-0001-9790-0395

  • Steven Ian Higgins

    Contributed equally to this work with: Edmund Carl February, Steven Ian Higgins

    Affiliations Department of Botany, University of Otago, PO Box 56, Dunedin 9054, New Zealand, Biodiversity and Climate Research Centre (BiK-F), Senckenberg Gesellschaft für Naturforschung, Senckenberganlage 25, 60325, Frankfurt am Main, Germany

Rapid Leaf Deployment Strategies in a Deciduous Savanna

  • Edmund Carl February, 
  • Steven Ian Higgins
PLOS
x

Abstract

Deciduous plants avoid the costs of maintaining leaves in the unfavourable season, but carry the costs of constructing new leaves every year. Deciduousness is therefore expected in ecological situations with pronounced seasonality and low costs of leaf construction. In our study system, a seasonally dry tropical savanna, many trees are deciduous, suggesting that leaf construction costs must be low. Previous studies have, however, shown that nitrogen is limiting in this system, suggesting that leaf construction costs are high. Here we examine this conundrum using a time series of soil moisture availability, leaf phenology and nitrogen distribution in the tree canopy to illustrate how trees resorb nitrogen before leaf abscission and use stored reserves of nitrogen and carbon to construct new leaves at the onset of the growing season. Our results show that trees deployed leaves shortly before and in anticipation of the first rains with its associated pulse of nitrogen mineralisation. Our results also show that trees rapidly constructed a full canopy of leaves within two weeks of the first rains. We detected an increase in leaf nitrogen content that corresponded with the first rains and with the movement of nitrogen to more distal branches, suggesting that stored nitrogen reserves are used to construct leaves. Furthermore the stable carbon isotope ratios (δ13C) of these leaves suggest the use of stored carbon for leaf construction. Our findings suggest that the early deployment of leaves using stored nitrogen and carbon reserves is a strategy that is integrally linked with the onset of the first rains. This strategy may confer a competitive advantage over species that deploy leaves at or after the onset of the rains.

Introduction

Savanna is characterised by a discontinuous mosaic of trees with a grass understory [1]. Coupled with high diurnal temperatures, rainfall in savanna is distinctly seasonal with very pronounced wet and dry cycles. During the dry season rainfall may be extremely low or entirely absent for a few to several weeks. As a result, soil water availability is recognised as one of the key factors constraining maximum woody cover at more arid sites, whereas at less arid sites it is proposed that woody cover is not limited by moisture but by fire and herbivory [2, 3]. Research in the Kruger National Park in South Africa using stable water isotope ratios has shown that although riparian evergreen trees such as Philenoptera violaceae use deep water deciduous savanna trees such as Colophospermum mopane are reliant on regular rainfall [4]. This reliance on rainfall, combined with the rainfall seasonality, effects a decline in soil moisture in the dry season resulting in a suspension of growth in trees. In many tree species the dry season suspension of growth is additionally associated with deciduousness [58].

Several factors have been hypothesised to promote deciduousness in seasonally dry environments. For example, deciduous trees do not have to carry the costs of leaf respiration or contend with herbivory in the dry season [9, 10]. The liability is that deciduous trees have to construct new leaves, which is expensive both in nitrogen and carbon. The leaves of deciduous trees typically have higher nitrogen content, photosynthesis and specific leaf area than those of evergreen trees. This allows them to fix sufficient amounts of carbon in the relatively short growing period [911]. Several recent studies have suggested that nitrogen is a limiting resource in both South African and South American savannas [1214]. It has also been shown that nitrogen is highest in the top 10 cm of the soil and that the majority of roots are also in this layer of the soil [15, 16], that is plants in these systems allocate their roots preferentially in areas where nitrogen concentrations are highest. Furthermore, this nitrogen is available only briefly to plants, supplied as a pulse at the start of the growing season. [1719].

The high nitrogen demands of deciduousness, the low average nitrogen availability, and the seasonal pattern of nitrogen and water availability in the study system imply that nitrogen conservation may be important. Deciduous trees can conserve nitrogen through nutrient resorption or the withdrawal of nutrients from senescing leaves and the storage of these nutrients in woody parts [2023]. Many savanna trees deploy new leaves, just before the rains, using these stored resources [2426]. All deciduous trees at our study site may flush either before or with the first rains. Some species such as Sclerocarya birrea flush well before the rains (early flushing) while others such as Combretum apiculatum flush shortly before or with the first rains (late flushing) [6]. While the early flushing trees clearly pre-empt rainfall the late flushing trees do so to a lesser extent [6]. In this study we examine whether trees resorb nitrogen from leaves at the end of the growing season and whether stored carbon and nitrogen resources are used to deploy new leaves at the onset of the growing season. We do this to elucidate how trees can afford deciduousness in a nitrogen limited ecosystem.

Methods

Declaration

The study did not involve any humans or animals and permission to do the research in the Kruger National Park was granted by South African National Parks Scientific Services in the Kruger National Park. Our data are available on the national park web site at; http://dataknp.sanparks.org/sanparks/metacat/judithk.110818.7/sanparks

Study Site

The study site is located in the southern and south western part of the Kruger National Park, Mpumalanga Province, South Africa. Rainfall for the region is distinctly seasonal with hot, wet summers and cool, dry winters. The growing season starts in October and ends in April. Many trees are deciduous and grasses inactive in the dry season. Mean annual maximum and minimum temperatures at Skukuza (24°59'44.88"S 31°35'30.84"E) for the 30 years from 1981 to 2010 was 29.9°C and 14.7°C with mean annual precipitation 590 mm (South African Weather Service. WB42, Climate Statistics). Temperatures range between a maximum of 32.6°C in January to a minimum of 6.2°C in June.

The vegetation for the area is classified as granite lowveld by Mucina and Rutherford [27] with Sclerocarya birrea (A.Rich) Hochst, Combretum apiculatum (Sond.), and Terminalia sericea (Burch. Ex DC), common trees. Grasses are dominated by Hyperthelia dissoluta (Nees ex Steud) Clayton and Elionurus argenteus Nees. Soil depth is variable but these nutrient poor granite soils are seldom deeper than 1.5 m to bedrock in our study area [15, 28]. The basement geology for the area is comprised of granite, gneiss and migmatite (Nelspruit granite suite, [29]). These basement rocks weather into distinct catenal sequences with sandy soils on the uplands and clayey soils in the bottomlands [27]. Soil nitrogen for the region is 0.52 mg g-1 soil and phosphorous (Bray II) is 3.23 mg kg-1 soil [12].

Rainfall, Soil moisture and Leaf Phenology

Weather stations were located at (1) 25°4'21.71"S 31°32'48.50"E and approximately 5 kms from site 1, (2) 25°6'39.60"S 31°30'54.00"E and approximately 14 kms from site 2, (3) 25°5'27.60"S 31°22'51.60"E. These weather stations included a Decagon Devices EC-5 soil moisture probe (Pullman, WA, USA) which was inserted 5 m away from the weather station horizontally into the soil at a depth of 20 cm while rainfall and temperature were recorded using a Rain-O-Matic tipping bucket rain gauge (Pronamic, Ringkøping, Denmark) and a Vaisala HMP50 temperature sensor (Campbell Scientific, Logan, UT, USA). These instruments were controlled with a CR200 data logger (Campbell Scientific Inc., Logan, UT, USA), which was powered by a rechargeable lead acid battery connected to a voltage-regulated solar panel. For rainfall and soil moisture we average the values for the stations at our three study sites.

At each of these three study sites we also tagged three Terminalia sericea and three Combretum apiculatum trees. For each individual tree, an index of leaf canopy fullness was estimated [8]. Two observers scored the percentage of leaves relative to the potential canopy area of leaves for each study tree every second week from 6th January 2009 to 6th January 2010. Scores were as follows: 0 represented no leaves, 1<2%, 2 between 2% and 25%, 3 between 26% and 50%, 4 between 51% and 75% and 5 >75% of the total canopy. The number reached is firstly an independent estimate and then a consensus between the two observers as to an appropriate value for each tree. This consensus score was recorded.

Nutrient resorption

For each study tree described in the previous section we measured how nitrogen content changed in leaves and in twigs of different branch orders as the growing season initiated and terminated. We randomly collected ten fully expanded fresh new leaves from each of these study individuals at the end of each month at the end (March to June 2009) and beginning (Oct, Nov, Dec 2009) of the growing season. At the same time we also collected a 4 cm long twig sample from each of the first three branching orders using a pair of secateurs. We used a power drill to collect the sample from the 4th branching order. We denoted these branching orders analogous to how stream orders are denoted; order-one being the terminal branch, order two being the point below where two order-one branches join, order-three being where two order-two branches join and so on.

Both leaf and wood samples were dried to a constant weight at 70°C in a forced convection oven (Scientific, series 200 oven, South Africa) and the bark and pith removed from the twig before grinding to a fine powder using a Retsch MM200 ball bearing mill (Retsch, Haan, Germany). Stable carbon isotope ratios relative to PDB (δ13C) and percentage nitrogen were then determined on these samples using a Thermo Finnigan Delta plus XP Mass Spectrometer coupled to a Thermo Finnigan Flash EA1112 Elemental Analyser with automatic sampler (Thermo Electron Corporation, Milan Italy).

Statistical analysis

The change in leaf nitrogen content was analysed as a linear regression against time (in months from the start of the growing season). The twig nitrogen content (N) are analysed using a linear model,

In this model time counts from 3 to 12 (March to December) and the quadratic term is used to describe the non-linearity of how nitrogen changes with time for some twig orders. Separate slope and intercept parameters are estimated for each twig order. Tree is treated as a random effect. The parameters of the model (S1 Code Listing) were estimated using JAGS [30]. JAGS is a general purpose software for fitting hierarchical Bayesian models using Markov Chain Monte Carlo (MCMC) simulation. The response variable was assumed to be normally distributed and un-informed priors were used.

Results

Rainfall Soil moisture and Leaf Phenology

As rainfall decreased from April to October (Fig 1) both tree species gradually lost their leaves. For T. sericea this reduction in leaf canopy area was faster than for C. apiculatum but by the middle of October 2009 both species had lost all their leaves (Fig 1). As soil moisture increased with the first rainfall at the beginning of November both species exhibited an immediate and rapid increase in leaf area. There were no leaves on our T. sericea study trees on the 21st of October and only a few leaves on two of our C.apiculatum, (0 mm of rain in the preceding 2 weeks) on the 5th of November (12.4 mm of rain) there were a few leaves on all trees and by 17th November (48 mm of rain) between 50 and 75% of total potential canopy (Fig 1). Rainfall at our study site from the 1 st September to 21st October was 2.3 mm ranging in amount from 0.1 mm to 1 mm. Mean maximum temperature for this time was 30°C while minimum temperatures were 15°C for September and 17°C for October.

thumbnail
Fig 1. Time series of daily rainfall, volumetric soil water content and tree canopy fullness index for Combretum apiculatum and Terminalia sericea.

https://doi.org/10.1371/journal.pone.0157833.g001

Nutrient resorption

For both T. sericea and C. apiculatum leaf percentage nitrogen decreased linearly with time from the beginning of the wet season (Nov, Dec) through to the middle of the dry season in June (Fig 2, Table 1). The rate of this decrease was similar for the two species: the 95% credible intervals of the slopes of these regressions spanned -0.21 to -0.16 for C. apiculatum and -0.18 to -0.12 for T. sericea (Table 1). The intercepts were higher for C. apiculatum than for T. sericea indicating that C. apiculatum had higher leaf nitrogen contents.

thumbnail
Fig 2. The percentage leaf nitrogen per unit dry mass against time since the start of the growing season (November to June) for Combretum apiculatum and Terminalia sericea.

No estimate was available for October because no leaves were present. Grey bars are T. sericea and open bars C. apiculatum. Values are means ± 1 SE.

https://doi.org/10.1371/journal.pone.0157833.g002

thumbnail
Table 1. Slope (% nitrogen per month) and intercept (% nitrogen) of a linear regression of leaf nitrogen content against time since the start of the growing season (November to June) for Combretum apiculatum and Terminalia sericea.

https://doi.org/10.1371/journal.pone.0157833.t001

At the end the growing season (March to April) nitrogen content increased in all branch orders (orders 1 to 4), consistent with resorption of nitrogen into deeper branch orders (Figs 3 and 4). At the start of the growing season (November to December) nitrogen content decreased in deeper (orders 2,3,4) branch orders, but increased in branch order 1; changes that are consistent with a movement of nitrogen from basal twigs to terminal twigs (Figs 3 and 4). All change slopes reported differed from zero except for the changes in the deeper branch orders at the start of the growing season for T. sericea. The slopes shown in Fig 4 were derived from the fitted polynomial models and their full conditional posterior distributions were computed (S1 Code Listing). Concomitant with this, the first fully expanded leaves of the growing season (November) have δ13C values that are enriched relative to any other month of the year (P<0.001). For C. apiculatum these values are -25.7 ± 0.23‰ for November with a range from-27.3 ± 0.2 to -28.0 ± 0.3‰ for the rest of the year (Fig 5). For T. sericea values are -23.4 ± 0.3‰ for November with a range from -26.6 ± 0.1‰ to -28.4 ± 0.2‰ for the rest of the year (Fig 5).

thumbnail
Fig 3. Seasonal variation of nitrogen content for twigs of different branching orders (order 1 terminal, order 4 more basal) of T. sericea (top) and C. apiculatum (bottom).

Values are means ± 1 SE.

https://doi.org/10.1371/journal.pone.0157833.g003

thumbnail
Fig 4. Slope of the relationship between twig nitrogen content and time at the end (March to April) and start (November to December) of the growing season for C. apiculatum and T. sericea.

The bars indicate the estimated slopes and the error bars indicate the 95% credible intervals of the estimates.

https://doi.org/10.1371/journal.pone.0157833.g004

thumbnail
Fig 5. Leaf δ13C values against time since the start of the growing season (November to June) showing the relatively enriched 13C values in November relative to other times of the year.

Values are means ± 1 SE.

https://doi.org/10.1371/journal.pone.0157833.g005

Discussion

The study system has a dry and cool unfavourable season and a wet, warm growing season resulting in a strong seasonality in leaf phenology [11, 25]. Our study species were completely deciduous, with no leaves on the trees at the end of the dry season. The patterns of canopy leaf fullness were correlated with rainfall, for example both species show an increase in canopy volume in response to a late rainfall event in May. While leaf fall may be closely aligned with rainfall amount several studies have suggested that leaf emergence may be decoupled from soil moisture [6, 11, 25]. Using a multi-spectral camera to track the normalized difference vegetation index (NDVI) of individual canopies Higgins et al [6] detected that leaf emergence pre-empted rainfall at the same study site. Although the Higgins et al [6] study identified individual tree canopies they could not assign species names to these canopies. Our results however show that only two of our study C. apiculatum trees had some leaves before the rains while there were no leaves on the T. sericea until after the rainfall event of the 30th October. This rain event at the end of October triggered bud-break and rapid leaf growth as we observed an extremely rapid expansion of the canopy for both species in response to the first rains from a few leaves on the trees at the end of October to a full canopy within two weeks of the first rains. This canopy response coincides with the flush of plant available nitrogen that becomes accessible when the soil is rewetted with the first rains at the end of the dry season [1719].

The new leaves produced at the beginning of the growing season were enriched in the heavier 13C isotope relative to other times of the year. Several studies have shown that leaves derived from stored reserves have δ13C values that are more enriched than leaves derived from recent photosynthesis [31, 32]. The reason for this is that part of these reserves is comprised of non-structural carbohydrates which have a more enriched (up to 4 ‰) δ13C value than those derived from recent photosynthesis [33, 34]. The depletion we observed in leaf δ13C values as the growing season progressed, for both of our study species, is consistent with the carbohydrates in these leaves being replaced by assimilates derived from more recent photosynthesis [32]. These results suggest that our study trees are using stored carbon reserves in the construction of new leaves.

In low nutrient environments such as at our study site trees may resorb nutrients from the leaves prior to abscission [13, 22]. Such a strategy divorces a plant from a reliance on a continuous and current nutrient supply. Our results show that nitrogen is transported from the leaves to the more proximal branches as leaf abscission progresses. At the onset of the next growing season this nitrogen is then remobilized to more terminal branches. This mobilisation of stored nitrogen and carbon allowed our study species to deploy new leaves and construct a full leaf canopy within two weeks of the first rainfall event. We therefore propose that trees in our study system can afford the cost of deciduousness by withdrawing and conserving both nitrogen and carbon.

In this study we use the two most common tree species at our study site and found that both follow the same strategy. That is, both species flush their leaves early synchronous with rainfall and are able to rapidly construct a full canopy after that. It may be argued that the primary function of rapid leaf deployment is to satiate browsing herbivores, which may be particularly beneficial at the beginning of the growing season when browse availability is limited [35]. If this were true selection might favour cheating where individual trees that did not invest in early-deployment would deploy leaves later once herbivores are satiated. An alternative is that the rapid construction of a full canopy of leaves allows plants to set-up transpiration gradients that, through mass flow from the soil to the roots, maximises uptake when nitrogen becomes available in a pulse at the beginning of the growing season [13, 36, 37]. In a previous study at this site we detected a large pulse of nitrogen with the first rains [19] and this relationship between rainfall and nitrogen mineralization has also been demonstrated in a bimodal rainfall savanna in Senegal [38]. The rapid construction of a full canopy of newly developing leaves may additionally set-up source sink gradients to initiate root development [39, 40].

Our findings suggest that the deciduous strategy of savanna trees is tightly linked to the seasonal pattern of nitrogen availability and that early deployment of leaves using stored nitrogen and carbon reserves is integral to this strategy. Future studies should examine whether the strategy of rapid leaf deployment allows trees to compete more effectively for soil nitrogen than a later flushing strategy.

Supporting Information

S1 Code Listing. JAGS Code listing for the statistical model used to generate the parameters shown in the main manuscript Fig 4.

"

https://doi.org/10.1371/journal.pone.0157833.s001

(BUG)

Acknowledgments

We are grateful to H. Combrink, A. Mashele, M. Mashele, P. Ndlovu, and P. Vimercati for fieldwork and the South African Weather Service for supplying us with the weather data. The authors have no conflict of interest to declare.

Author Contributions

Conceived and designed the experiments: ECF SIH. Performed the experiments: ECF SIH. Analyzed the data: ECF SIH. Contributed reagents/materials/analysis tools: ECF SIH. Wrote the paper: ECF SIH.

References

  1. 1. Frost P, Menaut JC, Walker B, Medina E, Solbrigo T. Responses of savannas to stress and disturbance. International Union of Biological Sciences Special Issue 10 Paris, France1986.
  2. 2. Higgins SI, Scheiter S, Sankaran M. The stability of African savannas: insights from the indirect estimation of the parameters of a dynamic model. Ecology. 2010;91(6):1682–92. pmid:20583710
  3. 3. Sankaran M, Hanan NP, Scholes RJ, Ratnam J, Augustine DJ, Cade BS, et al. Determinants of woody cover in African savannas. Nature. 2005;438(7069):846–9. ISI:000233777800049. pmid:16341012
  4. 4. February EC, Higgins SI, Newton R, West AG. Tree distribution on a steep environmental gradient in an arid savanna. Journal of Biogeography. 2007;34(2):270–8. ISI:000244179000009.
  5. 5. Archibald S, Scholes RJ. Leaf green-up in a semi-arid African savanna—separating tree and grass responses to environmental cues. Journal of Vegetation Science. 2007;18(4):583–94. ISI:000246899200015.
  6. 6. Higgins SI, Delgado-Cartay MD, February EC, Combrink HJ. Is there a temporal niche separation in the leaf phenology of savanna trees and grasses? J Biogeogr. 2011;38(11):2165–75.
  7. 7. Chidumayo EN. Climate and Phenology of Savanna Vegetation in Southern Africa. Journal of Vegetation Science. 2001;12(3):347–54.
  8. 8. Rossatto DR, Hoffmann WA, Franco AC. Differences in growth patterns between co-occurring forest and savanna trees affect the forest–savanna boundary. Functional Ecology. 2009;23(4):689–98.
  9. 9. Eamus D. Ecophysiological traits of deciduous and evergreen woody species in the seasonally dry tropics. Trends in Ecology & Evolution. 1999;14(1):11–6.
  10. 10. Givnish TJ. Adaptive significance of evergreen vs. deciduous leaves: solving the triple paradox. Silva Fennica. 2002;36(3):703–43.
  11. 11. Franco AC, Bustamante M, Caldas LS, Goldstein G, Meinzer FC, Kozovits AR, et al. Leaf functional traits of Neotropical savanna trees in relation to seasonal water deficit. Trees. 2005;19(3):326–35.
  12. 12. Craine JM, Morrow C, Stock WD. Nutrient concentration ratios and co-limitation in South African grasslands. New Phytologist. 2008;179:829–36. pmid:18537887
  13. 13. Kozovits A, Bustamante M, Garofalo C, Bucci S, Franco A, Goldstein G, et al. Nutrient resorption and patterns of litter production and decomposition in a Neotropical Savanna. Functional Ecology. 2007;21(6):1034–43.
  14. 14. van der Waal C, de Kroon H, de Boer WF, Heitkonig IMA, Skidmore AK, de Knegt HJ, et al. Water and nutrients alter herbaceous competitive effects on tree seedlings in a semi-arid savanna. Journal of Ecology. 2009;97(3):430–9. ISI:000265035400006.
  15. 15. February EC, Higgins SI. The distribution of tree and grass roots in savannas in relation to soil nitrogen and water. South African Journal of Botany. 2010;76(3):517–23. ISI:000279986300013.
  16. 16. February EC, Cook GD, Richards AE. Root dynamics influence tree–grass coexistence in an Australian savanna. Austral Ecology. 2013;38:66–75.
  17. 17. Williams CA, Hanan N, Scholes RJ, Kutsch W. Complexity in water and carbon dioxide fluxes following rain pulses in an African savanna. Oecologia. 2009;161(3):469–80. pmid:19582479
  18. 18. Jarvis P, Rey A, Petsikos C, Wingate L, Rayment M, Pereira J, et al. Drying and wetting of Mediterranean soils stimulates decomposition and carbon dioxide emission: the “Birch effect”. Tree physiology. 2007;27(7):929–40. pmid:17403645
  19. 19. Higgins SI, Keretetse M, February EC. Feedback of trees on nitrogen mineralization to restrict the advance of trees in C4 savannahs. Biology letters. 2015;11(8):20150572. pmid:26268994
  20. 20. Chapin FS, Kedrowski RA. Seasonal Changes in Nitrogen and Phosphorus Fractions and Autumn Retranslocation in Evergreen and Deciduous Taiga Trees. Ecology. 1983;64(2):376–91.
  21. 21. Pallardy SG. Nitrogen Metabolism. In: Kramer PJ, Kozlowski TT, editors. Physiology of Woody Plants. Third edition ed. New York: Academic Press; 2008. p. 233–54.
  22. 22. Ratnam J, Sankaran M, Hanan NP, Grant RC, Zambatis N. Nutrient Resorption Patterns of Plant Functional Groups in a Tropical Savanna: Variation and Functional Significance. Oecologia. 2008;157(1):141–51. pmid:18488252
  23. 23. Santa Regina I, Rico M, Rapp M, Gallego HA. Seasonal Variation in Nutrient Concentration in Leaves and Branches of Quercus pyrenaica. Journal of Vegetation Science. 1997;8(5):651–4.
  24. 24. Scholes RJ, Walker BJ, editors. An African Savanna. Synthesis of the Nylsvley study. Cambridge UK: Cambridge University Press; 1993.
  25. 25. Williams RJ, Myers BA, Muller WJ, Duff GA, Eamus D. Leaf phenology of woody species in a north Australian tropical savanna. Journal of Ecology. 1997;78:2542–58.
  26. 26. Chapin FS. The mineral nutrition of wild plants. Annual Review of Ecology and Systematics. 1980:233–60.
  27. 27. Mucina L, Rutherford MC, editors. The vegetation of South Africa, Lesotho and Swaziland. Pretoria: South African National Biodiversity Institute 2006.
  28. 28. Coetsee C, Jacobs S, Govender N. An overview of nitrogen cycling in a semiarid savanna: some implications for management and conservation in a large African park. Environmental management. 2012;49(2):387–402. pmid:22057696
  29. 29. Barton JM, Bristow JW, Venter FJ. A summary of the Precambrian granitoid rocks of the Kruger National Park Koedo. 1986;29:39–44.
  30. 30. Plummer M, editor JAGS: A program for analysis of Bayesian graphical models using Gibbs sampling. Proceedings of the 3rd International Workshop on Distributed Statistical Computing (DSC 2003) March; 2003.
  31. 31. Damesin C, Rambal S, Joffre R. Seasonal and annual changes in leaf d13C in two co-occurring Mediterranean oaks: relations to leaf growth and drought progression. Functional Ecology. 1998;12(5):778–85.
  32. 32. Helle G, Schleser GH. Beyond CO2-fixation by Rubisco—an interpretation of 13C/12C variations in tree rings from novel intra-seasonal studies on broad-leaf trees. Plant, Cell & Environment. 2004;27(3):367–80.
  33. 33. Badeck FW, Tcherkez G, Nogués S, Piel C, Ghashghaie J. Post-photosynthetic fractionation of stable carbon isotopes between plant organs—a widespread phenomenon. Rapid Communications in Mass Spectrometry. 2005;19(11):1381–91. pmid:15880634
  34. 34. Brugnoli E, Hubick KT, von Caemmerer S, Wong SC, Farquhar GD. Correlation between the carbon isotope discrimination in leaf starch and sugars of C3 plants and the ratio of intercellular and atmospheric partial pressures of carbon dioxide. Plant Physiology. 1988;88(4):1418–24. pmid:16666476
  35. 35. Murali K, Sukumar R. Leaf flushing phenology and herbivory in a tropical dry deciduous forest, southern India. Oecologia. 1993;94(1):114–9.
  36. 36. Barber SA. A diffusion and mass-flow concept of soil nutrient availability. Soil Science. 1962;93(1):39–49.
  37. 37. Cramer MD, Hawkins H, Verboom GA. The importance of nutritional regulation of plant water flux. Oecologia. 2009;161(1):15–24. pmid:19449035
  38. 38. Bernhard-Reversat F. Biogeochemical cycle of nitrogen in a semi-arid savanna. Oikos. 1982:321–32.
  39. 39. Davies WJ, Zhang J. Root signals and the regulation of growth and development of plants in drying soil. Annual Review of Plant Biology. 1991;42(1):55–76.
  40. 40. Buttrose M. The effect of reducing leaf area on the growth of roots, stems and berries of Gordo grapevines. Vitis. 1966;5:455–64.