Estimating root production at different

(1)

Estimating root production at different elevations on a salt marsh using soil cores,

ingrowth cores and stable isotopes.

J.A. Vonk Rijksuniversiteit Groningen Supervisors:

J.P. Bakker (RuG) A.C. Bockelmann (RuG) T.J. Bouma (NIOO-CEMO) February 2001

(2)

Abstract

This

study was focused on a biological and a methodological question on root production and turnover. During the growth season, from May to September,

ingrowth cores and soil cores were harvested for root production. Ingrowth cores on a high and low salt-marsh site were compared for root production. The high site had a significant (p<O.OO1) higher production of roots, but a significant (p<O.O5) lower rhizome production. Soil cores and ingrowth cores on the high site were compared,

because both methods have some disadvantages. No consistency was found in

previous studies in production using both methods. Before the first harvest, stable isotopes, 13C and '5N, were added in mid May, to calculate an independent value for root production and turnover. The roots from the harvested cores were ^{divided by} colour, for better tracing of the isotope during the growth season.

For all soil- and ingrowth cores a new core was placed after harvest. With these short term or re-growth cores root production during a month was measured. Re-allocation of nitrogen from the roots could be found if the ratio '3N:ö '5N increased or when the '5N in high in the re-growth core roots. However, no isotope data were available yet.

.D 2I

(3)

Introduction

Below ground dynamics

Root production and turnover are important in ecosystem functioning. These

belowground dynamics play an integral role in ecosystem processes (McClaugherty et a!. 1982; Conn and Day 1993). Understanding of fine root dynamics is necessary for addressing important issues on several levels of resolution in ecology (Aber et a!.

1985). Much is unknown for belowground production and -regulation in many ecosystems. The main factor that influences root turnover in grasslands and scrub is mean annual temperature, because maintenance respiration increases exponentially with temperature. Higher temperature leads to increasing nutrient mineralisation rates.

In grasslands and shrub there is a strong correlation between air and soil temperature, while wetland soil temperature is moderated by the site hydrology (Gill and Jackson 2000).

The extent to which nutrients are re-allocated prior to turnover is important for estimating nutrient cycling in ecosystems (Aerts et a!. 1989; Conn and Day 1993;

Eissenstat er al. 2000; Gordon and Jackson 2000). Plants are continuously losing nitrogen through mortality of plant parts (e.g. root turnover), herbivory, leaching from leaves, production of seeds and root exudation, but internal re-allocation is a source of nitrogen for the plant (Hopkinson and Schubauer 1984). In ecosystems with nitrogen as a limiting factor for plant production, such as salt marshes, it had been hypothised that conservative nutrient cycling will be favoured (Kiehl et a!. 1997; van Wijnen and Bakker 1999). For example in the salt marsh, a plant has to reduce nitrogen losses by translocation of nitrogen from senescing leaves to the roots (Berendse and Aerts 1987). Given that little, if any, re-translocation of nutrients occurs before fine-root senescence, root turnover is a strong nutrient sink for most plants (Gordon and Jackson 2000). Low root turnover rates (Conn and Day 1993) and increase root life span (Ryser 1996) seem to inhibit nutrient losses.

The proportional allocation of photosynthetically fixed carbon to the root and shoot system of a salt-marsh plant is an important element in the C-cycle of tidal salt marshes (Hemminga er a!. 1996). The belowground C-cycle of the salt marsh is crucial for the understanding of the total salt marsh- and the global C-cycle and the changes in these cycles. In particular, because salt marshes fix more carbon dioxide

(5)

annually than most natural ecosystems (Howes et a!. 1985). It is, however, still not clear whether the salt-marsh carbon flux acts as a sink or a source for carbon. The productivity of salt marshes as a whole depends on the proportion of the various species and stands (Cruz and Hackney 1977). At the global level, fine root production represents a large and relative unknown portion of the production and decomposition carbon balance (Aber et a!. 1985). Structural C is a major input in the soil organic matter. Quantification of this part of the C cycle is necessary in order to ^accurately model C and other nutrient cycles, and also has implications for C storage in relation to the global C cycle (Stewart and Metherell 1999). At the plant community level, changes in carbon and nitrogen allocation between roots and aboveground tissues are thought to affect competition between species and thus future community dynamics (Aber et al. 1985). Shifts in plant life form might also influence rates of root turnover in ecosystems, and so influencing carbon and nutrient cycling (Gill and Jackson 2000). In the light of elevated atmospheric CO2 changes vegetation structure and root turnover, the need for a total understanding of the C-cycle increases (see Fitter et a!.

1997; van Ginkel and Gorissen 1998; Eissenstat er a!. 2000; Norby and Jackson 2000;

Woodward and Osborne 2000).

Salt marsh

Tidal salt marshes are highly productive coastal ecosystems (Hemminga et a!. 1996).

The major contributors to this productivity are the vascular plants, although benthic and planktonic algae can also make a significant contribution (Bakker et al. 1993). In ungrazed Wadden-Sea salt marshes nitrogen is the limiting nutrient at the high and low elevation (Kiehl et a!. 1997; van Wijnen and Bakker 1999), while sedimentation in the lower salt marsh acts as a sink of nitrogen (Bakker et a!. 1993). The highest values of soil nitrogen content are found in the low, old marsh (van Wijnen and Bakker 2000). The nitrogen-mobility (ratio mineral N: total N) in the soil of the salt marsh is high (2.6 to 8.7 %) compared to that in inland ecosystems (Bakker et a!.

1993). The role of salt marshes in the nitrogen cycle is mainly one of nitrogen

processing. The oxidised forms, like nitrate, are generally exported, and reduced forms, like ammonia and dissolved organic nitrogen, are imported (Abd. Aziz and Nedwell 1986; Bakker et a!. 1993). De-nitrification rate is probably higher during the winter period and at low-marsh elevations, due to more flooding and therefore more anoxic conditions (van Wijnen and Bakker 1999). Salt marshes are thus interesting for

(6)

root turnover measurements because: (1) Roots and rhizomes represent a reservoir in the energy and material cycles in which considerable energy, carbon, and nutrients are stored in estuanne wetland communities (Schubauer and Hopkinson 1984). (2) Belowground production often outweighs aboveground production (Bakker et a!.

1993). (3) The same plant species lives at different elevations on the marsh and has to cope with different habitats concerning their root turnover. (4) The systemis one of the few relatively undisturbed natural habitats.

Estimating root production

Estimating root biomass and production still is a major problem. The practical difficulties involved in determining root and rhizome production is an obvious reason for the relative scarcity of data on this point (Hemminga et a!. 1996). The high variability in root turnover estimations might partly be explained by methodological constraints. The methods used to measure and calculate belowground biomass and production can play a strong impact on determining estimates of root turnover (Gill and Jackson 2000). Measuring belowground primary production in grasslands and marshes has been a particularly troubling problem (Neill 1992). Direct methods (like core methods) should be utilised when studies are being initiated on a new site.

Indirect methods (based on nutrient fluxes) are useful for those ecosystems where data are already available on root biomass and production and they are accompanied with sufficient data on the pools and fluxes of abiotic resources (Vogt et a!. 1998).

A common approach to direct root measurement is sequential sampling of root biomass. Root production is then estimated by increments of standing stocks of living

and dead roots (Aerts and de Caluwe 1992). This soil core method gives a

representation of fine root biomass at one point of time and is a common method to estimate root biomass and production. Disadvantages of the soil core method are: (1) The amount of time and labour, and the resultant financial costs, associated with the cleaning and sorting roots from the cores (Schubauer and Hopkinson 1984; Howes Ct a!. 1985; Vogt et a!. 1998; Makkonen and Helmisaari 1999). (2) The growth of fine roots cannot be assessed (Makkonen and Helmisaari 1999). (3) It does not fully account for the biomass turnover between the sampling intervals (Schubauer and Hopkinson 1984; Howes et a!. 1985). (4) The problem of deciding what is the best way of predicting fine root production after the root cores have been processed (Singh et al. 1984; Howes et al. 1985; Vogt et a!. 1986; Publicover and Vogt 1993).

(7)

With the ingrowth core method an intact soil core removed from the ground is replaced with an equivalent volume of root free soil from the site or with sand (Vogt et a!. 1998). The root free soil added back into the hole is enclosed within a plastic mesh that can be used to remove the cores after leaving them in the field for different time intervals. The subsequent growth of roots into this core is used to estimate fine root production in the field. This method was introduced by Flower-Ellis and Persson in 1980 and has been applied to both agricultural crops and forest ecosystems (Vogt et

a!.

1998). Disadvantages of the ingrowth core method are: (1) The problem to

physically and chemically reconstruct the root free soil environment so that similar root production is measured inside and outside the core (Neill 1992; Vogt et al. 1998).

(2) Is the initial root-free, homogenised soil gives a good approximation of root

growth that would normally occur in

the competitive, non-homogenised soil environment (Vogt et a!. 1998; Makkonen and Helmisaari 1999). (3) It is difficult to determine how root production in a root-free zone differs from one that already is occupied by roots and whether root free soil induces micro-sites of higher growth (St.

John er a!. 1983; Vogt et a!. 1998). If sand is used, the core method will provide a good estimate of relative growth rates but a low estimate of total growth (Valiela et a!.

1974). In contrast to the soil core method, the ingrowth core method avoids the high labour and high variability associated with sorting live and dead roots from local soil and allow a clear identification of the time period in which root growth occurred.

However, it does not eliminate the problem of simultaneous biomass production and loss. Additional problems of disturbance associated with coring, like penetration resistance inside the cores must be considered (Neill 1992).

The soil core method can be used for studying both the annual and seasonal biomass variations. For estimation of production, sampling should be done at short intervals.

The ingrowth core method is more suitable for estimating the potential of annual fine- root production and relative growth rates of roots between sites (Vogt et a!. ^1998;

Makkonen and Helmisaari 1999). A problem of all root methods is whether the approach underestimates or overestimates root production (Persson 1978; Singh et a!.

1984; Vogt et a!. 1986; Santantonio and Grace 1987; Aerts et a!. 1989; Publicover and Vogt 1993).

Because of all problems associated with coring methods, stable isotopes of carbon and

nitrogen (C'3 and N'5) can be added to both core methods. It

is assumed by Hemminga (1996) that (1) the brief exposure to '3C-enriched carbon dioxide and its

(8)

subsequent assimilation provides a representative picture of the pattern of carbon^flow in the plant over a longer time scale, and (2) the '3C-tracer is not selectively treated in different biochemical pathways. We think that the same assumptions can be made for N'5. It is assumed that in the root the '3C were mainly used in the cell wall, while the '5N were used for proteins in the cell content. So the carbon stayed in the cell wall and would not be re-allocated, while the nitrogen could easier be re-allocated, although this was not found in previous studies.

Pulse labelling may give an independent value of the production and turnover. Pulse labelling make the plant parts formed at the moment of the labelling traceable.

Labelled roots can be traced until senescence. The use of '3C and '5N for pulse labelling is advantageous as it is a stable isotope and precautions and regulations necessary for using radioisotopes are not needed. The problems of pulse labelling may be: (1) None or too little isotope is present in the roots. (2) The turnover is too slow to measure in the time interval of this study. (3) The difficulty to distinguish the decrease in label in the roots due to growth and due to turnover. Therefore it is important to label the plants at the moment of stable growth and senescing. Pulse labelling at the beginning of the season in both the soil core- and the ingrowth core sites can be used to better compare both methods.

Research questions

This study is focused on a biological and a methodological question. Two different

sites on the salt marsh were compared for root turnover and production using

ingrowth cores. To compare the soil core method with the ingrowth core method, both methods were used on the same site. Pulse labelling with '3C and '5N was added to all the sites in May to get a better perception on root growth at the sites. The research questions were: (1) What is the root production and root turnover at the high- and at the low site and is there a difference between the elevations on a salt marsh? (2)Is the ingrowth core method a suitable method for estimating root production and turnover in a natural site? (3) Does the labelling give a better perception of the turnover in soil- and ingrowth cores? (4) Does nitrogen re-allocation occur from the old roots to the new?

(9)

Hypothesis

(1) The environmental conditions at the low site are worse, e.g. more inundation ^and therefore less oxygen in the soil. The production is expected to be lower than on the high site. For turnover it is hard to predict a difference, because low production and a small standing biomass can give the same turnover as a high production and a large standing biomass. (2) The ingrowth core can not be a fully representative for the natural situation. This is quite diverse through years of environmental variations. The ingrowth core is on every site the same, and creates a small empty spot in the natural vegetation. For a root this is its entire environment. The influence of micro-site

differences between the soil- and ingrowth core can have a large impact. The

comparison between the advantages and disadvantages of the ingrowth cores can not be fully given until an independent value for root production and turnover is given. (3) The combination of the sequential cores and ingrowth cores, which are replaced by re- growth cores during the season, and the addition of the stable isotopes, '3C and '5N,

gives the experimental design a strong resolution for the estimation of root

production, turnover and live span. In other studies there is a great variance in the

numbers found for turnover and production because

^all

methods have some

disadvantages for estimating the production. (4) Nitrogen is always the limiting factor at the marsh. The plants have to be nitrogen conservative, but little re-allocation of nitrogen from senescing roots is found in previous studies. Root turnover costs are expensive in a nitrogen budget. So we can not say if there is a difference in re- allocation, if occurring, of nitrogen from senescing roots at the different elevations at this moment.

(10)

Material and Methods

Site description

The study area was situated on the salt marsh of 'The Oosterkwelder' at the^coastal Barrier Island of Schiermonnikoog, The Netherlands ⁽⁵³⁰ ^26'N, 5°28'E). Both sites, the high and the low, were just south of the 'Kobbeduin'. This area had been ungrazed by livestock since 1958 and is relatively undisturbed by man. Tidal range is 2.3 m and MHT is 1.00 m above Dutch Ordnance Level (NAP) (van Wijnen and Bakker 1999).

The soil consisted of a thin layer of clay on top of the indigenous sand layer. Tidal influence was low at the high site (table 1). The low site had a thicker clay layer due to more inundation per year. This thicker clay layer was associated with a higher denitrification rate in the soil (van Wijnen and Bakker 2000). At the study sites, Elymus athericus (specie names follow van der Heijden 1990) grew at different elevations and had spread more towards the lower marsh during the former years (A.C. Bockelmann, personal communication). This succession towards domination of one or few plant

species in various sites over the salt marsh zones could be a result of natural

eutrophication. Extra deposition of nitrogen through atmospheric deposition could accelerate this process (Bakker et al. 1993; 01ff et a!. 1997).

At both sites E. athericus was the dominating species. E. athericus had a high uptake of nitrogen and grew fast. The down slope distribution boundary was a result of competition, rather than of physiological limits (Bockelmann and Neuhaus 1999). On

the high sites E. athericus was dominant but Festuca rubra occurred in lower

abundances. The low site was almost only covered E. athericus. Later in the season

Table 1. Differences in the sites. (mean ± SE; n=20). Differences between sites shown in characters.

The height was relative to Mean High Tide (MHT). The inundation frequencies were derived from three flow-over loggers in the field at different heights (Bockelmann er a!. 2001 submitted). The N- pool of the soil was calculated assumed that every mm of clay deposition was equivalent to 1.9 gNm2 (van Wijnen and Bakker 2000). The higher number of inundation caused the thicker clay layer on the low site. The sites of the soil core and the high ingrowth core were largely comparable. The numbers of tillers were present on the labelled site (0 29 cm).

Site with treatment

Height to MHT

(cm)

Inundation (# year')

Clay layer thickness

(cm)

N-pool in clay (g N m2)

E. athericus tillers (#1 0 29 cm)

#E. athericus of total tillers

(%) Soilcore

high

74.2± 1.4

a

37.0± 1.7 a

9.1 ±0.5 a

172±9

a

144±7

a

95.2± 1.1

a

Ingrowth corehigh

75.1±2.0

a

36.8±2.6

a

10.1±0.5

a

191±10

a

112±7

b

84.0±3.5

b

Ingrowth corelow

38.9±0.6

b

137.4±4.0

b

16.5±0.5

b

313±9

b

148±10

a

99.1±0.4

C

(11)

Atriplex prostata grew between the E. athericus on both sites. For more details of the salt marsh area see 01ff et a!. (1997) and van Wijnen and Bakker (1999).

Cores

Table 2. Overview of the cores. All data from 2000. On two different sites two core methods were used. The ingrowth cores were placed mid February. All treatments were harvested from mid May till September, every two weeks two replicates per treatment. Every treatment core was directly after harvest replacedbya re-growth core. These were harvested a month after placing. Only of the lasttwo harvested of each treatment no re-growth core was placed. These places were directly filled up to restore the salt marsh.

Treatment Placing Harvest Re-growth_cores

first First last

Site Core date cores/

harvest

total first last placed harvest harvest

# # date date date date date

high Soil X 2 20 5-21 9-21 5-21 6-19 9-23

high ingrowth 2-23 2 20 5-20 9-20 5-20 6-19 9-23

low ingrowth 2-23 2 20 5-22 9-20 5-22 6-20 9-23

Three types of cores were used in this study; soil cores, ingrowth cores and re-growth cores (table 2). Because the soil cores and the high ingrowth cores were on the same high site under the same conditions (table 1), this is considered as one site with two treatments. All cores were 25 cm long and 6.5 cm in diameter. The soil cores were cores taken from undisturbed places at the high site. The ingrowth cores were tubes of plastic fence (length 25 cm, 0 6.5 cm and mesh width 1,0 cm), filled with root-free mud from the bottom of a creek. The mud was oxidised on the creek bank during a couple of days. The re-growth cores were the replacements of soil- and ingrowth cores during the growth season for a new and clean ingrowth core. The re-growth core was harvested a month after placing. The ingrowth cores were placed mid February 2000 in the salt marsh at a high site and a low site. So the treatments were at the high site soil cores and ingrowth cores, and at the low site ingrowth cores, which were all replaced by a re-growth core when the original core of the treatment was harvested (table 2). Around all the cores the vegetation was dominated by E. athericus. The

cores at the low site were placed about 1 meter next to Atriplex portulacoides-

vegetation, which is a typical low salt-marsh species.

Labelling

In mid May 2000, before the first harvest, the plots around 20 soil cores and 20

ingrowth cores at each site were incubated with '3C02

^using small Plexiglas greenhouses (height 50 cm, 0 29 cm, volume 33 1). The Plexiglas greenhouses were

(12)

placed around the core. No air change could occur between the greenhouse and the environment. 12.5 ml 0.500 M H2S04 (aq) was injected to a pot with 0.5 g NaH'3C03

(s) inside the greenhouse to release the '3C02. The plant took up the

^'3C02 photosynthetically during two hours with full sun or four hours with clouds. At the end of the '3C labelling 100 ml 3.33 mM '5N}{4Cl were added to the ground surface.

The next day the '5N labelling was repeated. It was not possible to inject the '5NH4CI-

solution into the soil, because debris stucked the needle. The '3C and the '5N were built in the plant, so there was a higher value of these isotopes in the parts of the plant that were formed at that moment. The label was followed during the season in the different root fractions. Assumed was that

four days

after incubations the translocation of carbon from the exposed leaves to the other plant parts was completed (Hemminga et a!. 1996). Five days after the incubation one young leave and one old leave were harvested at each treated plot, to estimate the amount of isotope taken up by the plant. From each treatment three cores were harvested to estimate the natural abundance of the stable isotope in the root fractions.

Harvest

10 days after the incubation the first two cores of each treatment were harvested; two soil-, two high ingrowth- and two low ingrowth cores. A re-growth core was placed back. Of each harvest place clay thickness and number of shoots were measured in the incubated plot (0 29 cm). The soil cores were split up in the upper 10 cm and the de lower part from 10 till 25 cm depth to make the washing and the splitting up of the roots less labour intense. The roots were rinsed over a 500 l.tm sieve. The fraction was divided in dead biomass and living roots. The dead biomass was divided in dead leave or aboveground biomass and debris. The living roots were subdivided by colour, because this was an indication of the age (vitality) of the roots (Aerts et al. 1989).

Young roots were white and after whitish and yellow they became brown and grey after which they decayed (see table 3 for the different fractions taken into account).

After the splitting the fractions were weighted, frozen, freeze dried for 3 days,

weighted again and crushed with a bullet crusher for the total C and N, '3C and '5N analyses with a mass photo-spectrometer (Stewart and Metherell 1999). This was repeated every two weeks during the growth season from May till September.

(13)

Table 3. Root classes by colour to root qualities. The roots and other material from all the cores were divided using these classes of colour and quality. For all the biomass data of the individual fractions, see appendix.

Fractions sorted:

Root classes by colour and other fractions

Root quality or fraction description Root fractions mentioned in this report (pooled data)

White Fresh formed roots.

White Thick white Fresh formed thick roots (0 >0.5mm).

Whitish Fresh formed roots, which already became soft and glassy.

Yellow/brown Standing roots, still fresh with none black tips and woody parts.

Yellow/brown Brown/grey Standing roots with black tips, woody parts,

not full function anymore. Most dying or dead root biomass.

Brown/grey Rhizome Used for vegetative reproduction to spread

the plant over a short distance.

Rhizome F. rubraroots Thin black roots, none E. athericus

Other species Other species roots ^{None E.} athericus orF. rubraroots, but all

other species.

Leaf Dead aboveground biomass present in the

core, like dead leaves and spikes. Dead biomass Debri All unidentifiable fine material, root pieces

of length <5 mm and thickness < ^0.2mm.

Data analysis Root production

Root production can be calculated from soil core data using the following equations:

(la) Dahlman and Kucera (1965): P (production) =maxbIOS — minbIO,S,.

This

max -

^mm method subtracts minimum measured biomass for any sampling data during the year from the maximum. The max-mm method assumes only that root production and mortality are asynchronous; that root production occurs in pulses of one to several months' duration and that mortality does not occur during this

period. Neill (1992) modified this equation by dividing biomass in roots and

rhizomes:

(ib) P =

^(LRhmax ^- ^LRhmin)

+ (LRX -

^LRIIUfl), with live rhizomes (LRh) and live roots (LR).

(2a) Singh and Yadava (1974): P =

6LT,

öLT>O,

with live total biomass (LT). The biomass increment between the sampling intervals,

if positive, together gives the production. Neill (1992) modified this

equation by dividing biomass in roots and rhizomes:

(2b) LRh

⁺ LR), LRh>O, LR>O.

(14)

Hansson and Steen (1984) used this equation (2a) on the increments of both biomass and necromass:

(2c) LT + NT),

LT>0, NT>0, necromass (NT), or the increment of necromass only öNT, öNT>O. Aerts et a!. (1989) defined this equation: ^If

(2d1) L>O and öD>O, P L +^5D;

(2d2) öL>0 and D<0, P =

(2d3) L<O and D>0, P =

L

⁺ ^öD or P =0 (when negative) and (2d4) öL<O and öD<0, P =0, with living roots (L) and dead roots (D).

Fairley and Alexander (1985) made a decision matrix in which the living and dead root biomass compartments were balanced.

This method

^is ^known ^as ^the compartmental flow method (Santantonio and Grace 1987; Vogt et al. 1989):

(3) P = B211 + M211 + Dt2t,,

with B211 the statistically significant change in live root biomass between time 1 and time 2; Mt2.tj the statistically significant change in dead root biomass between ti and t2; and Dt2t1 ^anestimate of root decay between ti and t2.

At this moment it is not possible to calculate the production in the soil cores using a max — ^mm approach, because we do not have the biomass data yearround. The necromass fluctuated much during the season and was not really reliable, because a lot of the fine debris (<500 tm) was rinsed trough the sieve and it was verydifficult to determine the origin, above- or belowground, of the dead material. Only the ^methods using summations of increments can be used, but due to Vogt et a!. (1998) it is clear that if non-significant differences between sampling dates are summed to ^determine root growth, root production will be over-estimated since random errors and biases

will accumulate when summing differences between sampling dates.

To calculate root production from the ingrowth- and re-growth cores data Neill ⁽¹⁹⁹²⁾ used the equations:

(4a) Sum of the biomass increments in the short-term cores (=^re-growth ^core).

(4b)Total biomass accumulation in the long-term cores (=ingrowth core).

(4c) Sum of the peak root and peak rhizome biomass in the long-term cores.

(4d) Total biomass in the long-term cores plus turnover losses. Turnover losses were calculated as the differences in biomass between the sum of the short-term cores^and the long-term cores at each interval, when the sum of the biomass in the ^short-term cores was greater than the corresponding long-term cores.

(15)

All these equations can be used.

Root turnover

Determining of root turnover, in literature, is done with the equations:

(5a) Dahiman and Kucera (1965): Turnover= annual growth total root mass

(5b) Gill and Jackson (2000): Turnover= annual belowground production maximum belowground standing crop with maximum root biomass = 0.45 ^* BNPP+ mean root biomass BNPP =below-ground net primary production

(5c) Aber er a!. (1985): Turnover= annual belowground production mean belowground standing crop (5d) Hendrik and Pregitzer (1983): Turnover= annual belowground production

minimum belowground standing crop Equations 5b, Sc and 5d were derived from the Dahlman and Kucera (1965)

calculation, In the Dahlman-Kucera model, an annual plant would have a turnover of 1.0 yr' if all of the roots that it produced were to die at the end of the growing season (Gill and Jackson 2000).

Our experimental setup does not provide in annual data on the belowground standing crop and therefore we can not estimate the max/mean/mm belowground standing crop. Consequently we can not estimate the root turnover using these equations. The turnover during the harvest period can be measured using the values of the ^ingrowth

core as the production and the values of the soil cores as

the max/mean/mm belowground standing crop.

(6) Root turnover = ^root produced in the ingrowth core

max/mm/mean belowground standing crop in the soil cores The problem of these calculations is that a comparison of growth into an ingrowth core (root free) is linked with the biomass values in the natural situation. This is never done in other studies. The measured belowground standing crop in this study could not contain yearround maximum and and/or minimum and could show a ^different mean value.

Root lifespan

The individual root lifespan can be estimated from the ingrowth core data.

(16)

(7) By extrapolating the data of first root growth into the core till the ^brown/grey fraction (senescing roots) appeared, the longevity of the root can be calculated. For the lifespan of the white roots:

(8) the growth speed in the re-growth cores with the standing white root biomass can be used. This biomass is almost constant and the re-growth speed also, so the time for the replacing of the white roots can be made. For the total standing biomass this ^is more difficult, because of fluctuations in the biomass.

Calculations on the stable isotope data

The lifespan and production of the white roots can be estimated from the difference in time between labelling and the decrease of label from this fraction, together with the biomass data of this fraction. For the lifespan of the white roots the equation:

(9) Lifespan = ti3 -

tia,eiiei, with tCI3=O the day '3C from the labelled and control fraction equals 0, and tiabelled the day the plot was incubated. Also a biomass correction must be taken into account. The production for the white roots iscalculated as:

(10) Pwhite = biomasswhjte / lifespanwhjte.

Not all places were incubated under exact the same conditions and with the same biomass. Therefore, from the data of the stable isotope analysis the relative amount of the isotope in the different fractions is taken, to correct the differences in labelling intensity. To calculate the relative amount of '3C in the different fractions we used the equation:

(11) / total '3Ccore) * 100%

For '5N the same equation was used.

From the ingrowth- and re-growth cores production and the soil core standing

biomass estimation is made for the turnover speed of the roots. As the '3C is expected to stay in the same roots, the fraction in which the labelled roots are present can ^be estimated per core. The turnover calculations can be compared with the analysis data to confirm the reliability of the turnover values derived from the different ^methods and calculations.

Re-allocation of nitrogen can be estimated by comparing the relative amounts of ^'3C and '5N in the fractions. If the ratio increased in the young roots or decreased in the old roots, nitrogen was re-allocated. A second approach is analysing the roots in the re-growth cores. If the ö or '5N of these roots differ significantly from the control roots, this is caused by re-allocation of nitrogen.

(17)

Results

Plots

The site used for the soil cores and the high ingrowth cores was the same and

considered as one site (table 1). The numbers of tillers on the plots at the high ^marsh differed for soil and ingrowth cores. The low site was 35 cm lower, but was flooded

100 times more a year. Most biomass was present in the upper part of the soil. For the sequential cores (25 cm length) 70 ± 1 % (n=23) of the total living biomass was present in the upper 10 cm. 20 ± 2 % of the total biomass in the soil cores were living roots.

Standing ^bionzass

Time

(date)

Fig. 1. Total standing biomass in the soil cores (depth 25 ^cm, 0 6.5 cm), pooled depths 0-10 cm and 10-25 cm. at the high salt marsh during the season (mean ± SE; n=2). ^The biomass in the cores fluctuated more within harvests than between harvests. For the relative distribution of the different fractions, seefig 2.

Standing biomass in the soil cores fluctuated during the season because of natural variance in the field (fig ^1). No seasonal trend was found.

The different fraction in which the cores were divided, showed variance, but the percentage of white roots was almost stable during the season (fig^2). The shift from the fresh roots (yellow/brown) to the class of the senescing roots (brown/grey) occurred just after the generative reproductive period. At this moment it is not possible to calculate production and turnover from these harvests, but the added stable isotope label, '3C and '5N, can give more details. This analysis can give a more

1.7

0.5

5/14 6/23 8/2 9/11

(18)

1.6

c 1.2

0.8

0.4

0.0

detailed view of the exact place of the label in the roots during the season, because of the dividing of the roots in different root classes.

100

25

0

5/14

Time (date)

Fig. 2. The relative distribution of the different root classes from the soil cores (depth 25 ^cm, 0 6.5

cm) at the high salt marsh during the season (mean ± SE; n=2). ^White: new formed young roots;

yellow/brown: older fresh roots; brown/grey: old and senescing roots. For the absolute weights of the harvests, see fig 1.

The standing rhizome biomass strongly decreased in June and gradually increased from August onwards (fig3).

6/23 ^8/2 ^9/11

5/14 6/23 8/2 9/11

Time (date)

Fig. 3. Rhizome biomass in the soil cores (depth 25 cm, 0 6.5 cm) at the high salt marsh during the season (mean ± SE; n=2).

In growth cores

The ingrowth cores showed a significantly (p<0.OOl) higher root production for the high then for the low site during the season (fig ^4). The rhizome production was significantly (p<O.O5) higher on the low than on the high site (fig ^5).

(19)

' 0.9

0.6 E

o.3

Fig. 5. Rhizome biomass in the ingrowth cores (length 25 cm, 0 6.5 cm, mesh width 1.0 cm) on the high- and low salt marsh (mean ± SE; n=2). The differences were significant (ANOVA, df 44, p<O.OS).

Table 4. The relative distribution of the different root fractions from the high (a) and the low (b) salt marsh for the ingrowth cores (length 25 cm, 0 6.5 cm, mesh width 1.0 cm) (mean ± SE; n=2). White:

new formed young roots; yellow/brown: older fresh roots; brown/grey: old and dying roots. For the absolute weights of the harvests, see fig 4.

High salt marsh Low salt marsh

Time (date)

Fractions (%) Time

(date)

Fractions (%)

white yellow/brown brown/grey white yellow/brown brown/grey

05/20 05/31 06/15 06/27 07/19 07/26 08/09 08/14 08/22 09/06 09/20

79.3±8.9 88.8±0.7 100.0 ± 0.0 78.7±21.0 97.8 ±0.3 35.6 ± 6.0 41.4±6.7 43.0 ± 0.8 56.4±4.5 33.6 ± 1.0 3.2 ± 0.7

20.7±8.9 3.8±3.8 0.0 ± 0.0 20.9±20.9

1.4 ±0.1 62.8 ± 6.2 55.6±5.2 55.1 ± 0.8 41.9±3.6 62.4 ± 1.4 83.2 ± 3.7

0.0±0.0 7.3±3.2

0.0 ± 0.0 0.4±0.1 0.9 ±0.2 1.6 ± 0.2 3.0± 1.5 1.9 ± 0.9 1.7 ±0.9 4.0 ± 0.4 13.6 ± 4.4

05/22 06/02 06/17 06/28 07/19 07/26 08/10 08/17 08/22 09/06 09/20

78.7±21.3 93.5±6.5 60.8 ± 37.9

86.8±6.0 94.2±0.4 52.3 ± 26.4 73.4±22.4 40.0 ± 8.3 54.2± 1.8 48.1 ± 11.0 22.7 ± 3.6

21.3±21.3 6.5±6.5

37.7 ± 37.7 11.1±3.9 3.1 ±3.1 46.1 ± 24.8 26.6±22.4 56.7 ± 8.6 45.8± 1.8 50.4 ± 9.5 60.8 ± 3.5

0.0±0.0 0.0±0.0

1.5 ± 0.2 2.1±2.1

2.7 ±2.7 1.6 ± 1.6

0.0±0.0

3.3 ± 1.8

0.0±0.0

1.5 ± 1.5 16.5 ± 7.0 1.5

1.0 C

I

^0.5

0.0

5/14 6/23 8/2 9/11

Time (date)

Fig. 4. Root biomass in the ingrowth cores (length 25 cm, 0 ^6.5 cm, mesh width 1.0 cm) on the high- and the low salt marsh during the season (mean ± SE; n=2). The differences were significant^(ANOVA, df 44, p<O.OOl). The production at the low site started later.

0.89

0.0 5/14

0.71

6/23 ^8/2 ^9/11

Time (date)

(20)

Spatial differences between the harvests made the roots classes in the ingrowth cores fluctuate strongly (table 4). The percentage of the white roots declined and the old and

senescing class increased at both sites near the end of the

growing season in

September.

Re-growth cores

The growth speed in the re-growth cores (length 25 cm, 0 6.5 cm and mesh width 1,0 cm) was significantly (p<O.Ol) higher at the high salt marsh than at the low salt marsh (fig ^6a). The plots at the high marsh, which had an ingrowth treatment before the re- growth core, were also significantly (p<0.O5) higher than the plots that had a ^{soil core} treatment before (fig^6b).

a ^b ^a ^b

6.0

4.0

2.0

0.0

Fig.6a and 6b. Growth speedper core in the re-growth cores of the different sites and pre-treatments during the season (mean ± SE; n=16). (6a) The high- and the low salt marsh differed significantly (ANOVA, df 45, p<O.O1). (6b) Re-growth cores at the high site with pre-treatment of soil cores and ingrowth cores differed significantly (ANOVA, df 45, p<O.O5).

Root production

Calculating root production from the soil core data is not possible at this moment, because no yearround biomass data was obtained in this study. Only the method using the increment of biomass is usable for quatification, but the biomass fluctuated too much for this method. Especially with the small number of replicates per harvest, the method gives unreliable results.

For the calculation of the production in the re-growth cores the mean re-growth speed is multiplied by the total time interval between the first placing and the last harvest.

For the total production in the ingrowth cores, the trend-line is calculated at the day of the last harvest. For, The production in the ingrowth and re-growth cores at the high

site ranged from 0.60 g tol.40 g for the roots and from 0.08 g to 0.30 g for the

rhizomes (table 5). At the low site this was respectively 0.12 g to 0.62 g and 0.18 g to

(21)

0.99 g. Because of the low number of replicates per harvest equation 4d ^{from Neill} (1992) is not reliable to use.

Table 5. Root and production calculations for four months during the growing season (half May ^till halfSeptember). Valuesfor root and rhizome production usin equations from Neill (1992).

Roots Rhizomes

Site High Low High Low

Equation ^(g/4 ^months) (g/4 months) (g/4 months) (g14 months)

(4a) re-growth cores 0.60 ^0.12 ^0.08 ^0.18 (fig 6a and 6b)

(4b) ingrowth cores ^0.95 ^0.35 ^0.15 ^0.49 (fig 4 and 5)

(4c) Max ingrowth cores 1.40 0.62 0.30 0.99 (fig 4 and 5)

Root turnover

At this moment it is not possible to calculate turnover with equations derived ^from Dahiman and Kucera (1965). Ingrowth core root production and the soil core standing biomass together showed a variance in turnover between 0.36 and 2.59 for the period from May till September (table 6).

Table 6. Turnover values for roots at the high site for the period May till September. The different production values derived from table⁵ were combined with a max, mm and mean standing biomass in the soil cores. No significant differences can be calculated.

Turnover values for the period May till Seutember Standing biomass

(fig 1) (g)

Production (P) (table5) method (g/4 month)

Turnover P/ standingbiomass

(4 month')

Mm:

0.54

(4a) 0.60 ^1.11

(4b) 0.95 1.76

(4c) 1.40 2.59

Mean:

1.07

(4a) 0.60 0.56

(4b) 0.95 0.89

(4c) 1.40 1.31

Max:

1.65

4a) 0.60 0.36

(4b) 0.95 0.58

(4c) 1.40 0.85

Root lifespan

The root lifespan or longevity, derived from the first root growth into the ingrowth core till the first dead roots (equation 7) for the high and low site are respectively 11.5

and 9.5 weeks. Using the standing white root biomass and the re-growth data

(equation 8), the white root longevity is 4 weeks.

Isotope data

The isotope data is not available yet, so no ca'culations can be made. The comparison of production and turnover between the soil- and the ingrowth cores can not be made at this moment. We only know that the first harvest contained a high value of 6 ^'3C.

(22)

For the ingrowth cores it is not possible to calculate the turnover value from the ö dilution. The isotope was added to the cores in mid May, when the root biomass was not in a steady state so the dilution of the label by growth is too large to distinguish from dilution by turnover (fig ^4).

From the dilution and shifts of the label through the fractions, production and turnover values can be estimated in the soil cores.

(23)

Discussion

Root seperation

The separation of partly decomposed dead roots from the organic soil matter was very difficult (Makkonen and Helmisaari 1999) and errors could arise in the process of making decisions whether to separate roots into live

and dead

^categories (Clemensson-Lindell 1994). So the total living root biomass percentage in the soil is probably higher than found here. The dividing of the roots in different classes based on the colour of the roots was difficult (Schubauer and Hopkinson 1984; Makkonen and Helmisaari 1999). The decline in the older fresh roots towards the senescing roots could be the result of higher temperatures during the summer.

Standing biomass

At the end of the growth season an increase in root biomass was expected, because during the die back process of the above ground plants parts towards autumn their biomass is partially mobilised and transferred to the roots (Steinke et a!. 1996). Also more than 50% of aboveground translocation is re-located to the roots at the end of the growing season (Hopkinson and Schubauer 1984). This was not seen in this experiment. We could have stopped the experiment too early in Autumn, but using few numbers of samples with soil cores could possibly produce rather biased results due to spatial heterogeneity of root distribution in the soil (Aerts and de Caluwe

1992).

The change in standing rhizome biomass during the season may be explained by the change from vegetative— to generative reproduction in June and the shift back in August. The seeds were full-grown end August, after which the biomass increased

again. The seed filling represents an enormous sink, which can not be met by

photosynthesis and demands remobilization of reserves stored in the roots which then degenerate (von Steinke et a!. 1996).

Site differences

The numbers of tillers on the plots at the high marsh differed for soil and ingrowth cores., because of the ingrowth cores placing early in the season and the labelling with

(24)

the greenhouses. The low site was 35 cm lower, but was flooded 100 times more a year. Hence the clay layer was thicker, but the environmental conditions wereharder.

Oxidised iron was present around the roots of the natural soil core at the low site when the ingrowth cores were placed in February. This clearly indicated anoxic conditions in the soil, which was oxidised by the roots. Anoxic soil present in the lower part of the ingrowth cores during the whole season indicated less oxygen in the soil during this period. More inundation and therefore the more anoxic conditions caused the lower root production on the low site in both ingrowth cores and re-growth cores. Higher values of soil nitrogen content were present in the low marsh (van Wijnen and Bakker 1999) and root production is also lower when there are more nutrients available in the soil (Valiela et al. 1976). A.C. Bockelmann (unpublished results) found a difference in shoot and leave production by E. athericus at the high and the low marsh. The high marsh E. athericus produced longer shoots and leaves, and more leaves due to more intra-specific competition. At the low marsh a higher seed production, higher recruitment and lower densities were found due to higher environmental stress. This explained the higher root production at the high site and the rhizome production at the low site.

Methodological differences

The significant difference between the re-growth speed in the plots with pre-treatment of soil cores and ingrowth cores at the high site could be caused by the conformation

of the plants to grow in a root free environment. This was not present in the

undisturbed sites from which the soil core was taken, so this can be considered as an artefact created by the ingrowth cores. The larger disturbance associated with the placing of cores and lag time of growth after the placing of the cores caused the great differences in values between the ingrowth- and re-growth core production. Neill (1992) also found more biomass in the long-term cores compared with the sum of the short-term cores indicated a disturbance effect associated with insertion of the cores.

Turnover and lifespan

Because the production values from table 5 differed much, the turnover numbers even have a larger difference (table 6). Until these numbers can be compared with an independent value for turnover, from the stable isotope data or the traditional used calculations derived from Dahlman and Kucera, this variation remains. Individual root

(25)

longevity is determined primarily by soil micro-site conditions (Pregitzer et ^{a!. 1993,} 2000), growing season length (Gill and Jackson 2000) and plant mineral nutrient conservation (Eissenstat et a!. 2000). The lifespan at the high site was longer than at

the low site. The low site contains more nitrogen and the lifespan of roots in

negatively correlated with nutrient availability (Ryser 1996)

(26)

Conclusions

(1) What is the root production and root turnover on the high- and on the low site and is there a difference between the elevations on a salt marsh?

Root production in the ingrowth cores at the high site was higher than at the low site as expected but rhizome production was higher at the low site than at the high site.

These differences were caused by different

strategies

to deal with the local

environmental conditions. The value for root production using different ^methods

showed a large variance, which can not be solved at this moment.

From the work of den Hengst et al. (1999, unpublished

results) we expected a turnover rate of about 13 weeks for E. athzericus. A lifespan was found in this study, for the high and the low site respectively 9.5 and 11.5, due to differences ^between sites and locations. However, turnover values could not be compared, because the turnover value of the ingrowth cores could not be made The labelling was too early for a steady state biomass in the cores. Therefore the dilution by growth is too large to distinguish turnover from growth dilution.

(2) Is the in growth core method a suitable method for estimating root production and turnover in a natural site?

At this moment it is not possible to compare both methods (soil cores and ingrowth cores), so no answer can be given. In other studies, however, the amount of root production estimated using ingrowth cores and sequential soil coring is not consistent within the same study site. Makkonen and Helmisaari (1999) found less production ⁱⁿ ingrowth cores compared to soil cores for Scots pine fine-root biomass in ^Finland.

Persson (1983) obtained the same results for pine root production in Sweden using ingrowth cores as when using soil cores, while Neill (1992) recorded much ^higher fine-root production estimates with ingrowth cores compared to soil cores in a ^prairie marsh in Canada. The spatial variation could be greater than the variation in time in all studies (Makkonen and Helmisaari 1999) with low number of replicates per harvest used.

Both methods have their advantages and disadvantages, but until the different root methods can be compared to some independently derived root biomass value obtained from total carbon budgets for systems, one method cannot be stated to be the best^and

(27)

the method of choice will be determined from researcher's personal preference, experiences, equipment, and/or finances (Vogt et al. 1998).

(3) Does the labelling give a better perception of the turnover in soil- and in growth cores?

The early labelling of the ingrowth cores makes the comparison of turnover at the high and the low site not possible. So turnover in both methods is not ^comparable with this analysis. At the high salt marsh the turnover values from the ^{soil core} method and the stable isotope data can be compared.

(4) Does nitrogen re-allocation occur from the old roots to the new?

Untill the data are available, nothing can be concluded.

Acknowledgement

I thank Jos van Soelen, Bas Koutstaal, Hester van Santen, Rene Eschen, Hiske van Duinen and the participants of the Vegetation Dynamics course 2000^forthe help with the fieldwork, Daan Bos and Dries Kuiper for practical remarks and information. And

last, but not least, the supervisors of this project, Jan Bakker, Anna Christina

Bockelmann and Tjeerd Bouma.

(28)

References

Abd. Aziz, S.A. and Nedwell, D.B. (1986). The nitrogen cycle of an east coast, UK, saltmarsh: II, Nitrogen fixation, nitrifIcation, denitrification, tidal exchange.

Estuarine, Coastal and Shelf Science 22: 689-704.

Aber, J.D., Melilo, J.M., Nadelhoffer, K.J., McClaugherty, C.A. and Pastor, J. (1985).

Fine root turnover in forest ecosystems in relation to quantity and form of nitrogen availability: a comparison of two methods. Oecologia 66: 3 17-321.

Aerts, R., Berendse, F., Klerk, N.M., and Bakker, C. (1989). Root production and root turnover in two dominant species of wet heathlands. Oecologia 81: 374-37 8.

Aerts, R., Bakker, C. and de Caluwe, H. (1992). Root turnover as determinant of the cycling of C, N, and F, in a dry heathland ecosystem. Biochemistry 15: 175- 190.

Bakker, J.P., de Leeuw, J., Dijkema, K.S., Leendertse, P.C., Prins, H.H.T. and

Rozema, J. (1993).

Salt marshes along the coast of The Netherlands.

Hydrobiologia 265: 73-95.

Berendse, F. and Aerts, R. (1987). Nitrogen-use-efficiency: a biologically meaningful definition? Functional Ecology 1: 293-296.

Bockelmann, A.C. and Neuhaus, R. (1999). Competitive exclusion of Elymus athericus from a high-stress habitat in a European salt marsh. Journal of

Ecology 87: 503-5 13.

Clemensson-Lindell, A. (1994). Triphenyltetrazolium chloride as an indicator of fine- root vitality and environmental stress in coniferous forest stands: Applications

and limitations. Plant and Soil 159: 297-300.

Conn, C.E. and Day, F.P. (1993). Belowground biomass patterns on a coastal barrier island in Virginia. Bulletin of the Torrey Botanical Club 120 (2): 121-127.

de la Cruz, A.A., and Hackney, CT. (1977). Energy value, elemental composition, and productivity of below ground biomass of a Juncus tidal marsh. Ecology 58: 1165-1 170.

Dahlman, R.C. and Kucera, C.L. (1965). Root productivity and turnover in native prairie. Ecology 46: 84-89.

(29)

Eissenstat, D.M., Wells, C.E., Yanai, R.D. and Whitbeck, J.L. (2000). Building roots in a changing environment: implications for root longevity. New Phytol. 147:

33-42.

Fairley, R.I. and Alexander, I.J. (1985). Methods of calculating fine root production in forests. In: Fitter, A.H. (Ed) Ecological interactions in soil (4: 37-42). Special Publication of the British Ecological Society NO.

Fitter, A.H., Graves, J.D., Wolfenden, J., Self, G.K., Brown, T.K., Bogie, D. and Mansfield, T.A. (1997). Root production and turnover and carbon budgets of two contrasting grasslands under ambient and elevated atmospheric carbon dioxide concentrations. New Phytol. 137: 247-255.

Gallagher, J.L. and Plumley, F.G. (1979). Underground biomass

profiles and productivity in Atlantic coastal marshes. American Journal of Botany ^66(2):

156- 16 1.

Gill, G.A. and Jackson, R.B. (2000). Global patterns of root turnover for terrestrial ecosystems. New Phytol. 147: 13-3 1.

van Ginkel, J.H. and Gorissen, A. (1998). In situ decomposition of grass roots as affected by elevated atmospheric carbon dioxide. Soil Sci. Soc. Am. J. 62:951- 958.

Gordon, W.S. and Jackson, R.B. (2000). Nutrient concentrations

in fine roots.

Ecology 8 1(1): 275-280.

Hansson, A.C. and Steen, E. (1984). Methods of calculating root production ^and nitrogen uptake in an annual crop. Swedish Journal of Agricultural Research

14: 191-200.

Hemminga, M.A., Huiskes, A.H.L., Steegstra, M. and van Soelen,

^J. ^(1996).

Assessment of carbon allocation and biomass production in a natural stand of the salt marsh plant Spartina anglica using 13C. Marine Ecology Progress Series 130: 169-178.

Hendrick, R.L. and Pregitzer, K.S. (1993). The dynamics of fine root length, biomass, and nitrogen content in two northern hardwood ecosystems. Can. J. For. Res.

23: 2507-2540.

Hopkinson, C.S. and Schubauer, J.P. (1984). Static and dynamic aspects of nitrogen cycling in the salt marsh graminoid Spartina al:errnflora. Ecology 65: 961- 969.

(30)

Howes, B.L., Dacey, J.W.H. and Teal, J.M. (1985). Annual carbon mineralizatoin and belowground production of Spartina alterniflora in a New England salt marsh.

Ecology 66(2): 595-605.

Kiehl, K., Esselink, P. and Bakker, J.P. (1997). Nutrient limitation and plant species composition in temperate salt marshes. Oecologia 111: 325-330.

Makkonen, K. and Helmisaari, H.S. (1999). Assessing fine-root production in a Scots pine stand — comparison of soil core and root ingrowth core methods. Plant and Soil 210: 43-50.

McClaugtherly, C.A., Aber, J.D. and Melilo, J.M. (1982). The role of fine roots in the organic matter and nitrogen budgets of two forest ecosystems. Ecology 63:

148 1-1490.

Neill, C. (1992). Comparison of soil coring and ingrowth methods for measuring belowground production. Ecology 73(5): 1918-1921.

Norby R.J. and Jackson, R.B. (2000). Root dynamics and global change: seeking an ecosystem perspective. New P/ytol. 147: 3-12.

01ff, H., de Leeuw, J., Bakker, J.P., Platerink, R.J., van Wijnen, H.J., and de Munck, W. (1997). Vegetation succession and herbivory in a salt marsh: changes induced by sea level rise and silt deposition along an elevational gradient.

Journal of Ecology 85: 799-814.

Persson, H. (1978). Root dynammics in a young Scots pine stand in central Sweden.

OIKOS3O: 508-519.

Persson, H.A. (1983). The distribution and productivity of fine roots in boreal forests.

Plant and Soil 71: 87-101.

Pregitzer, K.S., Hendrick, R.L. and Fogel, R. (1993). The demography of fine roots in response to patches of water and nitrogen. New Phytol. 125: 575-580.

Pregitzer, K.S., King, J.S., Burton, A.J. and Brown, S.E. (2000). Responses of tree fine roots to temperature. New Phytol. 147: 105-115.

Publicover, D.A. and Vogt, K.A. (1993). A comparison of methods for estimating forest fine root production with respect to sources of error. Can. J. For. Res.

23: 1179-1186.

Ryser, P. (1996). The importance of tissue density for growth and life span of leaves and roots: a comparison of five ecologically contrasting grasses. Functional Ecology 10: 7 17-723.

(31)

Santantonio, D. and Grace, J.C. (1987). Estimating fine-root production and turnover from biomass and decomposition data: a compartment-flow model. Can J.

For. Res. 17: 900-908.

Schubauer, J.P. and Hopkinson, C.S. (1984). Above- and belowground emergent macrophyte production and turnover in a coastal marsh ecosystem, Georgia.

Limnol. Oceanogr. 29(5): 1052-1065.

Singh, J.S. and Yavada, P.S. (1974). Seasonal variation in composition, plant biomass and net primary productivity of a tropical grassland at Kurukshetra, India.

Ecological Monographs 44: 35 1-376.

Singh, J.S., Lauenroth, W.K., Hunt, H.W. and Swift, D.M. (1984). Bias and random errors in estimators of net root production: a simulation approach. Ecology 65(6): 1760-1764.

von Steinke, W., Willert, D.J. and Austenfeld, F.A. (1996). Root dynamics in a saltmarsh over three consecutive years. Plant and Soil 185: 265-269.

Stewart, D.P.C. and Metherell, A.K. (1999). Carbon ('3C) uptake and allocation in pasture plants following field pulse labelling. Plant and Soil 210: 6 1-73.

St. John, T.V., Coleman, D.C. and Reid, C.P.P. (1983). Growth and spatial

distribution of nutrient-absorbing organs: selective exploitation of soil heterogeneity. Plant and Soil 71: 487-493.

Valiela, I.,

Teal, J.M. and Persson, N.Y. (1976). Production and dynamics of

experimentally enriched salt marsh vegetation: Belowground biomass. Limnol.

Oceanogr. 21(2): 245-252.

Vogt, K.A., Grier, C.C., Gower, S.T., Sprugel, D.G. and Vogt, D.J. (1986).

Overestimation of net root production: A real or imaginary problem? Ecology 67(2): 577-579.

Vogt, K.A., Vogt, D.J., Moore, E.E. and Sprugel, D.G. (1989). Methodological considerations in measuring biomass, production, respiration and nutrient resorption for tree roots in natural ecosystems. In: Torrey, J.G., and Winship, L.J. (pp 217-232).Applications of continuous and steady-state methods to root biology. Kluwer Academic Press, Dordrecht.

Vogt, K. A., Vogt, D.J. and Bloomfield, J. (1998). Analysis of some direct and indirect methods for estimating root biomass and production of forests at an ecosystem level. Plant and Soil 200: 71-89.

(32)

van Wijnen, H.J. and Bakker, J.P. (1997). Nitrogen accumulation and plant ^species replacement in three salt-marsh systems in the

Wadden Sea. Journal of

Coastal Conservation 3: 19-26.

van Wijnen, H.J. and Bakker, J.P. (1999). Nitrogen and phosphorus limitation in a coastal barrier salt marsh: the implication for vegetation succession. Journal of Ecology 87: 265-272.

van Wijnen, H.J. and Bakker, J.P. (2000). Annual nitrogen budgets of a temperate coastal barrier salt-marsh system along a productivity gradient at low and high marsh elevation. Perspectives in plant ecology, evolution and systematics 3(2):

128- 141.

Woodward, F.!. and Osborne, C.P. (2000). The representation of root processes in models addressing the responses of vegetation to global change. New Phytol.

147: 223-232.

Estimating root production at different