Nitrogen fixation by Elaeagnus angustifolia in the reclamation of degraded croplands of Central Asia

Nitrogen fixation by Elaeagnus angustifolia in the reclamation

of degraded croplands of Central Asia

ASIA KHAMZINA,1,2 JOHN P.A. LAMERS1 and PAUL L.G. VLEK1

1 Center for Development Research (ZEF), Walter Flex Street 3, 53113 Bonn, Germany

2 Corresponding author ([email protected])

Received July 22, 2008; accepted February 15, 2009

Summary Extensive degradation of irrigated croplands,

due to increasing soil salinity and depletion of soil

nutrient stocks, is a major problem in Central Asia

(CA), one of the largest irrigated areas in the world. To

assess the potential for improving the productive capacity

of degraded lands by afforestation, we examined N2

fixation of Elaeagnus angustifolia L. in mixed plantations

with non-fixing Populus euphratica Oliv. and Ulmus

pumila L. Fixation of N2 was quantified by the 15N

natural abundance technique based on both foliar

and whole-plant sampling during five consecutive

growing seasons. Despite elevated root-zone soil

salinity (6–10 dS m�1) and deficiency in plant-available

P (4–15 mg kg�1), N2 fixation (%Ndfa) increased from

an initial value of 20% to almost 100% over 5 years.

Within each growing season, %Ndfa steadily increased

and peaked in the fall. Annual N2 fixation, determined

using foliar d15N, initially averaged 0.02 Mg ha�1,

peaked at 0.5 Mg ha�1 during the next 2 years and

thereafter stabilized at 0.3 Mg ha�1. Estimates based on

whole-plant d15N were <10% lower than those based on

foliar d15N. The increase in plant-available soil N was

significantly higher in E. angustifolia plots than in

P. euphratica and U. pumila plots. Increases in the

concentrations of organic C (19%), total N (21%) and

plant-available P (74%) in the soil were significant

irrespective of tree species. This improvement in soil

fertility is further evidence that afforestation with mixed-

species plantations can be a sustainable land use option

for the degraded irrigated croplands in CA.

Keywords: afforestation, 15N natural abundance technique,P. euphratica, salinity, soil fertility, U. pumila, Uzbekistan.

Introduction

Afforestation has proven to be effective in re-vegetating sal-

ine landscapes, reducing elevated groundwater tables and

thus mitigating dryland salinization in irrigated land use

systems (Marcar and Crawford 2004). Planting N2-fixing

tree species on nutrient-exhausted fallow lands or inter-

planting them with crop species is a common silvicultural

practice that utilizes the ability of N2-fixers to replenish

nutrient stocks and increase the productivity of agrofor-

estry and silvopastoral systems, particularly in tropical

regions (Dawson 1986, Dommergues 1987, Danso et al.

1992, Paschke 1997, Galiana et al. 2004). To date, however,

few studies have examined biological N2 fixation and its

contribution to soil fertility in Central Asia (CA). This

region, one of the largest irrigated areas in the world, could

benefit greatly from improved forest fallows given that

extensive areas have been withdrawn from agricultural

use because of soil salinization and land degradation

(World Bank for Reconstruction and Development 2002,

Saigal 2003).

Afforestation, as an alternative land use system for

degraded agricultural areas, can be further enhanced by

planting both N2-fixing and non-fixing fast-growing multi-

purpose tree species that provide fuelwood, high-protein live-

stock fodder, edible fruits and, in the long term, timber

(Galiana et al. 2004, Khamzina et al. 2006, Lamers and

Khamzina 2008,Lamers et al. 2008). The actinorhizal species

Elaeagnus angustifolia L., native to CA, is salt-tolerant and

useful for fruit, fuelwood, gum, leaf fodder, nectar andhoney

production, medicinal purposes and amenity planting

(Hayes 1976, Ayaz et al. 1999, Katz and Shafroth 2003,

Khamzina et al. 2006). BecauseE. angustifolia is able to grow

in a wide range of climatic and soil conditions and particu-

larly on disturbed sites, it has been widely propagated as a

wildlife, windbreak, highway and reclamation species (Uresk

and Yamamoto 1994, Heinze and Liebmann 1998, Forestry

Compendium 2000). In riparian ecosystems, E. angustifolia

often colonizes aggressively, displacing native vegetation

and hence is regarded as a noxious exotic weed, e.g., in the

United States (Katz and Shafroth 2003). However, mixing

valuable tree crops such as Juglans nigra L. with N2-fixing

Elaeagnus umbellata Thunb. can improve growth perfor-

manceofJ. nigraas a result of increased soilNmineralization

rates (Friedrich and Dawson 1984, Ponder 1988, Paschke

1997). Further research is needed to determine if this bio-

fertilizer quality of Elaeagnus holds true on degraded

� The Author 2009. Published by Oxford University Press. All rights reserved.For Permissions, please email: [email protected]

Tree Physiology Page 1 of 10

doi:10.1093/treephys/tpp017

Tree Physiology Advance Access published March 20, 2009 by guest on O

ctober 1, 2014http://treephys.oxfordjournals.org/

Dow

nloaded from

http://treephys.oxfordjournals.org/

agricultural lands in CAwhere N2 fixation may be restricted

because of salinity, P-deficiency or insufficient soil water con-

tent – conditions known to inhibit the Frankia symbiosis

(Danso et al. 1992, Galiana et al. 2004).

The dearth of information on N2 fixation of actinorhizal

tree species compared with crop and pasture legume fixa-

tion has been attributed to the practical difficulties of

repeated whole-plant sampling because of the large size

and year-to-year variations in N2 fixation of trees (Dawson

1986, Boddey et al. 2000, Galiana et al. 2004, Gehring et al.

2005). Given these constraints, most published results have

been based on seedlings during their first 1–2 years of

growth (Galiana et al. 2004). Longer-term studies are

needed to determine temporal changes in the reliance of

trees on biologically fixed N for nutrition.

Among the various approaches available, the 15N natural

abundance method has proven to be the most suitable for

quantifying N2 fixation in tree plantations and agroforestry

systems, subject to certain requirements being met, includ-

ing the selection of appropriate reference plants and adjust-

ment for isotopic fractionation within the N2-fixer

(Domenach et al. 1989, Boddey et al. 2000, Peoples et al.

2002, Galiana et al. 2004, Unkovich et al. 2008). We used

the 15N natural abundance method to (1) evaluate seasonal

and year-to-year changes in N2 fixation by E. angustifolia

planted on a degraded agricultural site; (2) quantify the

contribution of N2 fixation to the plant–soil system; and

(3) assess changes in soil fertility following afforestation.

Materials and methods

Study region

The study was conducted on 2 ha of degraded agricultural

land in an experimental tree plantation located in the

Khorezm Region of Uzbekistan (41�650 N and 60�620 E,altitude 102 m a.s.l.). This region is a part of the Amu

Darya River lowlands within the transition zone of the

Karakum and Kyzylkum deserts and belongs to the CA

semi-desert zone that is characterized by an extremely con-

tinental climate. Mean annual rainfall of 100 mm falls

mostly outside the growing season and is greatly exceeded

by the potential evapotranspiration of 900–1000 mm.

Following decades of intense cultivation using unsustain-

able irrigation practices (Saigal 2003) about 15–20% of

the irrigated land in Khorezm is currently considered to

have low or no suitability for cropping. This study is part

of an international research program to assess the viability

of alternative uses of this land, particularly conversion to

multipurpose tree plantations (Martius et al. 2004).

Afforestation site

In 2003, a long-term experiment was initiated based on a

two factorial split-plot design comprising three tree species

(split factor) and three irrigation techniques (main factor)

as previously described (Khamzina et al. 2008). The tree

species were the actinorhizal Russian olive (E. angustifolia;

Elaeagnaceae) and two non-fixers, Euphrates poplar

(Populus euphratica Oliv.; Salicaceae) and Siberian elm

(Ulmus pumila L.; Ulmaceae). Elaeagnus angustifolia and

P. euphratica naturally co-exist in the poplar-dominated

riparian forest of the Amu Darya River. Both species have

also been planted within the nearby irrigated agricultural

area. Ulmus pumila was introduced to CA during the last

century and has been widely grown as a windbreak in the

irrigated area ever since.

The tree species were arranged in 36 pure, completely

randomized plots (12 plots per species), spaced 3.5 m from

each other. Each experimental plot, 105 m2 in size,

was planted in April 2003 with 70 one-year-old saplings

at a spacing of 1 m · 1.75 m, resulting in a density of

5714 trees ha�1. Given the dense planting, the plots devel-

oped closed canopies during the first growing season. The

trees received 80–160 mm year�1 of water applied either

by the furrow or drip irrigation method during the first

two growing seasons in 2003 and 2004. Because the

irrigation treatments did not have a significant effect on15N signals or N concentrations, this factor was omitted

from all our analyses.

Irrigation ceased from 2004 onwards and trees were left to

rely on the shallow groundwater table that averaged 1.3m in

2003 and gradually dropped to 1.9 m in 2007. Throughout

this period, the mean root-zone water content decreased

but at all times was close to field capacity, which ranged

from 27% to 38% within the soil profile. Groundwater

salinity ranged between 0.7 and 4.8 dS m�1 during the

2003–2007 observation period (Khamzina et al. 2008). Soils

at the experimental site are old irrigated meadow alluvial

soils with a predominantly silt loamy texture and a slightly

alkaline reaction. During 2003–2007, the degree of soil salin-

ity increased in the experimental plots from moderate-

to-strong as evidenced by the electrical conductivity (EC)

measurements. Soil fertility was initially poor because of

low organic carbon (C) and nitrogen (N), phosphorus (P)

and potassium (K) concentrations (Table 1). The concentra-

tions of major nutrients in the groundwater, measured at the

end of the seasons 2006–2007, were low (Table 2).

Plant sampling and analyses

Initially, to estimate N2 fixation by E. angustifolia, about

200 g of fresh leaves from a representative tree of each spe-

cies was collected from six of the 12 species-specific plots

(i.e., 18 plots sampled in all) at the beginning, middle and

end of the growing seasons (i.e., May, July and October)

in 2003–2005. Thereafter, all 36 plots were sampled

monthly throughout the growing seasons in 2006 and

2007. Fully expanded leaves were collected from various

levels of the canopy and bulked by the tree to integrate

within-tree variations (Domenach et al. 1989). At the end

of 2006 and 2007 seasons, all plant fractions were sampled

2 KHAMZINA, LAMERS AND VLEK

TREE PHYSIOLOGY

by guest on October 1, 2014

http://treephys.oxfordjournals.org/D

ownloaded from


including leaves, stems, twigs, coarse roots, fine roots and

nodules. The samples were dried at 60 �C for 72 h, finely

ground and analyzed for N concentration and 15N natural

abundance with a mass spectrometer (ANCA–SL/20-20,

SerCon, UK), with 2–3 replicates per sample. Samples col-

lected during 2003–2005 were stored and all were analyzed

at the same time in 2005. Samples collected in 2006 and

2007 were analyzed at the end of the year in which they

were collected.

Fixation of N2 by E. angustifolia was quantified by the15N natural abundance technique, according to Shearer

and Kohl (1986):

%Ndfa ¼d15Nref � d15Nfixing

� �d15Nref � B� �

" #� 100; ð1Þ

where %Ndfa is the proportion of N derived from atmo-

spheric N2, d15Nref is the 15N natural abundance of the

reference (non-fixing) species P. euphratica or U. pumila,

d15Nfixing is the d15N of the N2-fixing E. angustifolia and

B is the 15N value of the same N2-fixing species grown

with N2 as the sole nitrogen source.

For each growing season, %Ndfa was calculated using a

weighted mean of d15N based on whole-plant samples at

the end of 2006 and 2007 and whole-plant B values previ-

ously determined for actinorhizal species (Domenach

et al. 1989, Tjepkema et al. 2000), ranging from �1.41 to

�2.0. These authors did not specify d15N for leaves and

their reported whole-plant d15N values were higher than

the foliar d15N values of our field-grown trees. Therefore,

to determine the %Ndfa based solely on foliar d15N we fol-

lowed the approach of Peoples et al. (2002) and used the

minimal field-observed foliar d15N (�2.53) as the B value.

The annual rate of N2 fixation (Mg ha�1 year�1) was cal-

culated based on the N content accumulated annually in

plant tissues and the estimated %Ndfa. For the years

2003–2005, the N concentrations and 15N abundance in

the fractions other than leaves were assumed to be similar

to those measured in samples of these tissues collected in

2006 and 2007.

Biomass production

At the end of the growing season, in October 2003 and

2004, eight trees in each experimental plot were felled at

ground level and separated into stem, branches and foliage.

In 2005, four trees per plot were harvested and in subse-

quent years one tree per plot was used to determine dry

matter (DM) production. Coarse roots (Ø > 3 mm) were

excavated by hand to quantify belowground DM. In the

laboratory, roots were washed, separated from necromass

and grouped into several fractions according to root diam-

eter. Fine roots (Ø < 3 mm) and nodules were sampled at

each plot from four points around a tree (between trees in a

row and between rows) with a root auger, down to a depth

of 0.9 m. High soil water contents precluded sampling of

deeper horizons. For determination of DM, all plant parts

were dried at 103 �C to constant mass.

Soil and groundwater sampling and analyses

Before planting, soil texture and bulk density were

determined in six soil profiles throughout the field. Soil

was differentiated into seven size classes according to

Kachinsky (1965) and results converted to the Food and

Agriculture Organization (FAO) texture classification. At

the end of each growing season, soil (in 0.2-m layers down

to 1 m depth) and groundwater were sampled from all

Table 1. Electrical conductivity (EC, dS m�1), volumetric water content (h, %) and pH in the upper 1 m soil layer as well as organic C,

total N (g kg�1), nitrate, available P and exchangeable K (mg kg�1) concentrations in the upper 0.2-m soil layer in 2003–2007. The

P values were derived from an ANOVA of all data, with the level of significance set at P < 0.05. Values are the means of the soil

concentrations under the three species. Within a row, means followed by different letters differ significantly according to the Post Hoc

Tukey test. Abbreviation: ns, not significant.

Parameter 2003 2004 2005 2006 2007 P value

EC 6.2 c 7.8 b 6.7 c 9.6 a 9.8 a < 0.0001

h 37 a 34 b 33 c 31 d 31 d < 0.0001

pH (in water) 8 a 8 a 8 a 8 a 8 a ns

Organic C 7.8 b 8.1 ab – – 9.3 a < 0.05

Total N 0.52 b 0.56 ab – 0.57 ab 0.63 a < 0.01

N–NO3 – – – 15.4 25.5 0.01

P2O5 3.8 b 5.6 b – 11.6 a 14.8 a < 0.0001

K2O 157 a 149 a – 151 a 181 a ns

Table 2. Concentrations (mg l�1) of nitrogen (N), phosphorus

(P) and potassium (K) in groundwater at the end of growing

season in October 2006 and 2007 (mean ± standard error).

Elements 2006 2007

N 19 ± 5 1.4 ± 0.1

P 0.4 ± 0.1 0.3 ± 0.1

K 10 ± 1 12 ± 1

NITROGEN FIXATION BY TREES IN CENTRAL ASIA 3

TREE PHYSIOLOGY ONLINE at http://www.treephys.oxfordjournals.org



ownloaded from


experimental plots for nutrient analyses. Organic C was

analyzed according to the method of Tyurin (1975) based

on acidification of C with chromic anhydride in sulfuric

acid and titration. Total nitrogen (N) concentration was

analyzed by the Kjeldahl method. Nitrates were analyzed

colorimetrically by reaction with disulfophenolic acid in

alkali, according to Granval and Lajoux (Durinina and

Egorov 1998). Available phosphorus (P) was measured col-

orimetrically in an ammonium carbonate extract.

Exchangeable potassium (K) concentration was determined

by flame photometry.

Samples for the analyses of soil EC and water content

were collected with an auger every 15–30 days in all plots,

0.3 m distant from a tree stem, down to 1 m depth in 0.2-m

layers. Electrical conductivity of the soil water suspen-

sion (EC1:1) was measured with an EC meter. The EC1:1

values were converted to ECe based on the relationship

ECe = 3.6EC1:1 recommended by Shirokova et al.

(2000). Soil water content (h) was measured gravimetri-

cally (103 �C) and converted to volumetric values based

on soil bulk density. Soil water contents at field capacity

and wilting point were determined by the Richards

pressure membrane method (Klute 1986) for the six soil

profiles.

Statistical analyses

All data were checked for normal distribution and normal-

ized by logarithmic transformation when necessary. Data

were subjected to analysis of variance (ANOVA), which

was performed by applying the general linear model proce-

dure with SAS Version 9.1 software (SAS Institute, Cary,

NC), to determine significance of species, month, year

and plant fraction factors, with the accepted level of signif-

icance set at P < 0.05. The Tukey Post Hoc test was used

to compare the individual treatment means.

Results

15N natural abundance

The 15N abundance values of N2-fixing E. angustifolia were

predominantly negative and significantly different from

those of the two reference species, irrespective of tree age,

plant fraction or time of the growing season (Figure 1

and Table 3). In contrast, d15N values of U. pumila and

P. euphratica were positive and similar except in 2003 and

in spring of 2004–2005. The annual changes in 15N

abundance were significant in E. angustifolia, which initially

showed positive values turning to increasingly negative

values in older trees. Each year, a distinct seasonal pattern

was observed in all species with d15N gradually decreasing

during the period from April to October (Figure 1).

The nodules of E. angustifolia measured in October 2006

and 2007 were more depleted in 15N than the other tree

fractions analyzed (Table 3). Among the other tree

fractions and across all species combined, foliar 15N abun-

dance was lowest, whereas d15N in the rest of the fractions

did not differ significantly. In E. angustifolia and P. euphra-

tica, d15N in different fractions varied over the years and

according to species, but leaves always had the lowest

d15N values. The d15N of U. pumila did not vary signifi-

cantly among fractions (Table 3).

Nitrogen concentration

Nitrogen concentrations varied considerably among the

tree species (Figure 2), with highest values in E. angustifolia

at all times, followed by P. euphratica and U. pumila.

Similar to d15N, foliar N concentrations decreased during

each growing season until the leaf fall period. Foliar N con-

centrations were high in 2003–2004, subsequently decreased

and then stabilized in older trees (Figure 2). Among the dif-

ferent plant tissues examined in October 2006 and 2007,

nodules had the highest N concentrations exceeding the

concentrations in leaves and fine roots, whereas N concen-

trations in the woody above- and belowground fractions

were significantly lower and quite similar (Table 4).

In E. angustifolia, fine roots followed by nodules and then

leaves had much higher N concentrations than woody parts.

Among the woody fractions, stems had the lowest N concen-

trations. InU. pumila,N concentrationswere higher in leaves

than in fine roots, but both organs had higher N concentra-

tions than the woody fractions. In P. euphratica, N concen-

trations were highest in leaves. The relative concentrations

of N in the other fractions varied over time (Table 4).

Biomass production

Dry mass production differed significantly among the tree

species both overall and at different growth stages

(Figure 3). Elaeagnus angustifolia had the highest growth

rates in the early stages (2003–2005) but DM production

began to stabilize at around 100 Mg ha�1 in 2006. Dry

mass production was generally lower in U. pumila than in

E. angustifolia, but steadily increased throughout the obser-

vation period. Populus euphratica initially grew poorly,

hardly reaching 1 Mg ha�1 of total DM after the first sea-

son, but by 2007, P. euphratica trees were the largest among

the tree species accumulating about 120 Mg ha�1 in the

aboveground fractions alone (Figure 3).

The woody aboveground fraction contributed the largest

share to totalDMproduction for all species and ranged from

40% to 70% depending on tree age and species. Fine roots

were the smallest fraction comprising only 1–7% of the total

DM, although nodules of E. angustifolia contributed only

0.1–3.5% (data not shown). The leaf biomass share tended

to decrease over time in all species from 20–25% to only

6–15% in 2007, remaining the largest inP. euphratica.Ulmus

pumila had about 30–35% of biomass in its roots – signifi-

cantly more than the other species, especially E. angustifolia

which developed the smallest root fraction (Figure 3).


TREE PHYSIOLOGY



ownloaded from


Tissue N content

Given the high N concentrations in E. angustifolia tissues,

the largest amounts of N were stored in this species at all

times, even in 2006–2007 when its DM production was

overtaken by that of P. euphratica. The N content in

U. pumila steadily increased over time but was significantly

lower than in the other species from the third growing sea-

son onwards (Figure 4). The two reference species showed

significant differences in annual N accumulation in tissues

after the second growing season. In particular, P. euphra-

tica, which relied entirely on soil N, annually acquired more

N than the N2-fixing E. angustifolia in 2006–2007

(Figure 4).

Fixation of N2

In October 2003, the %Ndfa of E. angustifolia, estimated

using 15N natural abundance of U. pumila, significantly

deviated from the estimates based on d15N of P. euphratica.

With this exception, the choice of reference plants did not

influence the %Ndfa estimates significantly (Figure 5).

The initial %Ndfa estimated in October 2003 averaged

23% but increased to 82–88% over the next 3 years and

peaked at 100% in 2007. Each year, %Ndfa rose from

spring through summer and peaked in the fall. The %Ndfa

estimates based on foliar d15N were generally similar to

those based on whole-plant d15N. The only major difference

in %Ndfa estimated with the two methods was in October

2003 when d15N values were inconsistent between the refer-

ence plants (Figure 5).

Based on estimates of foliar d15N, the N content accumu-

lated annually in plant tissues through N2 fixation averaged

0.02 Mg ha�1 at the end of the first growing season,

peaked at 0.5 Mg ha�1 during the next 2 years and thereaf-

ter decreased and stabilized at 0.3 Mg ha�1 (Figure 6). The

use of %Ndfa based on whole-plant 15N natural abundance

resulted in estimates of annual N2 fixation 2–9% lower than

those based solely on foliar d15N, except for the first grow-

ing season when use of whole-plant weighted means yielded

—

—

—

Figure 1. Foliar 15N natural abundance inE. angustifolia, P. euphratica and U. pumila during2003–2007. The P values were derived from anANOVA of all data in the figure. Bars indicatestandard error of mean.

Table 3. Tissue d15N in E. angustifolia, P. euphratica and U. pumila at the end of the growing season in October 2006 and 2007. The

P values were derived from an ANOVA of the data for each year. Individual fraction means within each species and year and weighted

means within each year were compared by the Post Hoc Tukey test. Within a column, means followed by different lowercase letters are

significantly different at P < 0.05 between plant tissues. Within a row, weighted means followed by different uppercase letters are

significantly different at P < 0.05 between species. Abbreviation: ns, not significant.

2006 2007

E. angustifolia P. euphratica U. pumila E. angustifolia P. euphratica U. pumila

Nodules �2.29 b – – �2.48 c – –

Leaves �2.02 b 1.62 b 1.77 a �2.53 c 1.38 b 1.45 a

Fine roots �0.41 a 3.69 a 3.28 a �0.46 a 3.60 a 4.34 a

Stems �0.80 a 3.41 a 3.40 a �1.03 ab 2.09 ab 2.77 a

Branches �0.49 a 3.21 ab 3.49 a �1.64 bc 1.75 b 1.72 a

Coarse roots �0.66 a 2.41 ab 2.93 a �0.86 ab 2.38 ab 2.46 a

Weighted mean �0.78 B 2.94 A 3.16 A �1.32 B 1.96 A 2.35 A

P value

Species < 0.0001 < 0.0001

Fraction < 0.0001 < 0.0001

Species · fraction ns ns





ownloaded from


higher values compared with those based on foliar values

(Figure 6).

Effect on soil nutrient stocks

Over the 5-year observation period, exchangeableK concen-

trations were low and did not change significantly over time

in theupper0.2-msoil layer,whereas soil organicCandavail-

able P concentrations increased significantly but remained

relatively low (VSN 1985; Table 1). While the increases in

the concentration of these nutrients were unrelated to species

(results not shown), soil nitrate concentrations increased

most in the E. angustifolia plots (Table 5).

Discussion

Compliance with the requirements

of the 15N natural abundance technique

Accurate application of the 15N natural abundance tech-

nique requires selection of appropriate non-fixing reference

plants with similar rooting habits, phenological characteris-

tics and thus patterns of N uptake (Boddey et al. 2000).

These requirements were met in our study as evidenced

by the generally similar d15N of the two reference plants.

In part, this can be explained by the plantation’s shallow

groundwater table restricting vertical growth of roots,

which thus reside and acquire N at similar depths as the

N2-fixing species (Khamzina et al. 2008). The discrepancy

in d15N between the two reference plants at the beginning

of the study is explained by initial phenological differences

between the species (Danso et al. 1992, Boddey et al. 2000).

The species with a shorter life span, E. angustifolia and

P. euphratica, entered the vegetative stage earlier than

longer-living U. pumila, suggesting that the poplar was

initially a more reliable reference species.

In compliance with the requirements of the 15N natural

abundance method, the d15N of the reference plants, which

depended solely on soil N for nutrition, differed signifi-

cantly from that of the N2-fixer. Moreover, the d15N of

E. angustifoliawasmostly negative, a common phenomenon

observed in efficient N2-fixing systems (Boddey et al. 2000).

Figure 2. Foliar N concentrations in E. angustifolia,P. euphratica and U. pumila during five growingseasons. The P values were derived from anANOVA of all data in the figure. Bars indicatestandard error of mean.

Table 4. Tissue N concentration (%) in E. angustifolia, P. euphratica and U. pumila at the end of the growing season in October 2006

and 2007. The P values were derived from an ANOVA of the data for each year. Individual fraction means within each species and year

and weighted means within each year were compared with the Post Hoc Tukey test. Within a column, means followed by different

lowercase letters are significantly different at P < 0.05 between plant tissues. Within a row, weighted means followed by different

uppercase letters are significantly different at P < 0.05 between species.

2006 2007

E. angustifolia P. euphratica U. pumila E. angustifolia P. euphratica U. pumila

Nodules 2.8 ab – – 2.9 a – –

Leaves 2.3 b 1.4 a 1.3 a 2.2 b 1.4 a 1.4 a

Fine roots 3.0 a 0.6 b 0.5 b 2.4 b 0.9 b 0.7 b

Stems 1.1 c 0.6 b 0.4 c 0.7 d 0.2 d 0.3 c

Branches 1.3 c 0.4 b 0.3 c 1.1 c 0.5 cd 0.5 bc

Coarse roots 0.8 c 0.4 b 0.4 c 1.3 c 0.6 c 0.3 c

Weighted mean 1.3 A 0.6 B 0.4 C 1.1 A 0.5 B 0.4 B

P value

Species < 0.0001 < 0.0001

Fraction < 0.0001 < 0.0001

Species · fraction < 0.0001 < 0.0001


TREE PHYSIOLOGY



ownloaded from


In particular, nodules were depleted in 15N compared

with other fractions, which is in agreement with the

results of Tjepkema et al. (2000) on actinorhizal species

but contrasts with the results on many legumes showing

enrichment of root nodules (e.g., Yoneyama et al. 1984).

An adjustment for isotopic fractionation in N2-fixing spe-

cies is necessary when applying the 15N natural abundance

technique, particularly if the d15N of the N2-fixer is close to

or below zero (Danso et al. 1992, Doughton et al. 1992,

Boddey et al. 2000, Peoples et al. 2002). There are few pub-

lished B values for actinorhizal species and no information

was found for E. angustifolia. Applying a range of reported

B values for four actinorhizal species including Alnus glutin-

osa (L.) Gaertn., Casuarina cunninghamiana Miq., Datisca

glomerata (Presl.) Baill. and Myrica gale L. (Domenach

et al. 1989, Tjepkema et al. 2000), resulted in a 10–15% var-

iation in our %Ndfa estimates, suggesting that B values

specific to the Elaeagnus genus are required. The absence

of a reliable B value for the estimation of %Ndfa using

solely foliar d15N might explain the overestimation of

%Ndfa particularly in 2007, when the %Ndfa of 100%

was assumed. Nevertheless, the differences between the

estimates of N2 fixation based solely on foliar d15N values

and those based on whole-plant weighted means were

within acceptable limits.

The significant differences that developed over time in

annual tissue N accumulation among the study species,

along with the finding that the N2-fixer accumulated less

N than one of the reference plants, complicate application

of the total N difference method to accurately estimate

N2 fixation (Unkovich et al. 2008). Based on our findings,

we deem the 15N natural abundance method robust in

quantifying N2 fixation in perennials, in agreement with

the previous conclusions (Domenach et al. 1989, Boddey

et al. 2000, Galiana et al. 2004, Unkovich et al. 2008).

Temporal variation in N2 fixation

Temporal variability in 15N natural abundance, often

observed under field conditions, complicates studies with

perennial vegetation (Domenach et al. 1989, Boddey et al.

2000, Gehring et al. 2005). Seasonal variations in deciduous

trees can be integrated by sampling leaves at the end of the

growing season, before their senescence (Domenach et al.

1989). In support of this conclusion, we found that d15Ndecreased linearly towards the end of the season, when

N2 fixation was highest.

Figure 3. Biomass production ofE. angustifolia, P. euphratica andU. pumila at the end of the growingseasons in October 2003–2007. TheP values were derived from anANOVA of all data in the figure.Bars indicate standard error ofmean.

— —(Mg ha ) —(Mg ha )accumulation

Figure 4. Whole-plant N content and annual Naccumulation in E. angustifolia, P. euphratica andU. pumila estimated in October 2003–2007. Barsindicate standard error of mean.





ownloaded from


The apparent %Ndfa in perennial stands tends to decline

over time because ofN recycling fromaccumulated andmin-

eralized litter in mature plantations, particularly if the trees

are not pruned (van Kessel et al. 1994). In our study, the

%Ndfa, on the contrary, continuously increased over the

5 years, presumably because such a period was insufficient

for the litter to decay and significantly affect the 15N signals.

Quantity of biologically fixed N

Isotopic methods have proved to be robust when applied to

perennials although only a few studies have focused on

field-grown plants for periods longer than the first several

years of growth (Galiana et al. 2004, Gehring et al. 2005,

Unkovich et al. 2008). Several studies report between 9

and 119 kg ha�1 year�1 of N fixed with %Ndfa higher

than 75% for 2-year-old trees (Galiana et al. 2004). These

amounts of N fixed correspond to our results for 2003,

although we observed a %Ndfa of only 23%. The range

in annual rate of N2 fixation in our study was generally

higher than the 0.1–200 kg ha�1 year�1 reported in a

review by Paschke (1997) for actinorhizal species, examined

by the N accretion and acetylene reduction methods in the

western USA. Another review by Danso et al. (1992)

showed wide-ranging but generally much higher rates esti-

mated by the total N difference method of between 43

Table 5. Soil total N (g kg�1) and nitrate (N–NO3, mg kg�1) concentrations in the 0.2-m upper soil layer according to tree species at

the end of the growing season in October 2003–2007. The P values were derived from an ANOVA of all data in the table. Within a year,

the mean values of individual species were compared by a Post Hoc Tukey test. Within a column, means followed by different letters

are significantly different at P < 0.05 between species. Abbreviation: ns, not significant.

Parameter Species Year

2003 2004 2005 2006 2007

Total N E. angustifolia 0.53 a 0.60 a – 0.59 a 0.70 a

P. euphratica 0.51 a 0.56 a – 0.59 a 0.63 a

U. pumila 0.53 a 0.51 a – 0.53 a 0.56 a

P value

Species < 0.05

Year < 0.01

Species · year ns

N–NO3 E. angustifolia – – – 34.2 a 45.7 a

P. euphratica – – – 10.6 b 21.4 b

U. pumila – – – 14.3 b 27.1 ab

P value

Species < 0.0001

Year 0.01

Species · year ns

Figure 5. Nitrogen fixation (%Ndfa) of E. angustifolia deter-mined based on foliar and whole-plant 15N natural abundanceand a mean value for the reference plants P. euphratica andU. pumila during the growing seasons 2003–2007. Bars indicatestandard error of mean.

——

(Mg

ha)

Figure 6. Annual rate of N2 fixation by E. angustifolia estimatedbased on foliar and whole-plant 15N natural abundance and amean value for the reference plants P. euphratica and U. pumilain October of 2003–2007. Bars indicate standard error of mean.


TREE PHYSIOLOGY



ownloaded from


and 581 kg ha�1 year�1, which are comparable with our

values of 24–514 kg ha�1 year�1.

According to the tentative classification of Dommergues

(1987, 1995), species with an N2-fixing potential of

100–300 kg ha�1 year�1 are regarded as highly efficient.

Elaeagnus angustifolia would thus fit this category. The

N2 fixation rates we observed confirm that N2 fixation of

actinorhizal symbioses is not inferior to that of Rhizo-

bium-legume symbioses (Dawson 1986), and highlight the

efficient N2 fixation capacity of Elaeagnus spp., previously

reported by McNiel and Carpenter (1974) who compared

11 actinorhizal and legume species by the acetylene reduc-

tion method.

Drought, salinity, and P and micronutrient deficiency-

conditions that predominate in degraded soils in CA can

suppress N2 fixation by reducing plant growth or by

directly affecting the symbiosis (Dommergues 1995). At

our study site, soil water content was not a limiting factor

because of the shallow groundwater table that greatly con-

tributed to the high soil water content and transpiration

(Khamzina et al. 2009). The high soil salinity did not

adversely affect N2 fixation either, given the increasing

%Ndfa over time, implying high salt tolerance of both

the host plant and the Frankia strains naturally occurring

in the soil. The high rates of N2 fixation despite the P defi-

ciency may be attributed to the mycorrhizal associations on

roots stimulating efficient P uptake (Gardner 1986).

Perspectives for remediation of degraded soils

The ability for N2 fixation facilitated rapid initial growth

of E. angustifolia which, during the first 3 years,

outperformed the other species while investing the least bio-

mass in root development, because of low reliance on soil

N. Eventually, the DM production of E. angustifolia stabi-

lized approaching its genetic potential. The continuously

rising biomass production of the non-fixing P. euphratica

despite soil N limitations may be explained by the short-

term availability of N leached into the groundwater from

neighboring cropped fields, following their fertilization

(Kienzler, unpublished observations).

Soil organic C concentrations remained low after 5 years

of afforestation but the observed 19% increase is in accor-

dance with the results of a meta analysis (Guo and Gifford

2002) showing a mean 18% increase in soil C stocks follow-

ing conversion of cropland to tree plantation, based on 537

observations mostly from Australia, Brazil, New Zealand

and the USA. The efficient N2 fixation by Elaeagnus spp.

and the associated increase in plant-available N stocks

(Friedrich and Dawson 1984) were also confirmed in our

study, even though conducted under saline and P-deficient

conditions. In addition to soil C and N accumulation,

afforestation of degraded land can become important for

C sequestration in biomass and hence the emerging carbon

market.

In conclusion, the 15N natural abundance technique was

successfully used to estimate %Ndfa in E. angustifolia, at

least up to the age of 5 years, and can be further improved

by specifying the B value for the Elaeagnus genus. The tech-

nique was robust when based solely on foliar d15N. Nitro-

gen fixation by E. angustifolia persisted in saline and

P-deficient conditions and enriched plant-available soil N

concentrations, indicating that this species is highly suitable

for afforestation of degraded cropland in CA. The increases

in soil concentrations of organic C, total N and plant-avail-

able P were significant irrespective of tree species, thus all

species examined have the potential to serve as nutrient

pumps and improve soil fertility. We conclude that affores-

tation with mixed-species plantations can be a sustainable

land use option for the degraded irrigated croplands in

CA given the observed improvements in productivity of

the land and soil fertility following conversion to tree plan-

tations. The observed short-term effects are encouraging

but further simulation analyses are ongoing to assess the

long-term growth and environmental effects of afforestation

in CA.

Acknowledgments

The German Ministry for Education and Research (BMBF; Pro-

ject No. 0339970A), the Ministry for Schools, Science and

Research of the State of North Rhine-Westfalia financially sup-

ported this study. Additional funds from DAAD (German

Academic Exchange Service) and INTAS (International Associa-

tion for Promotion of Co-operation with Scientists from the

New Independent States of the Former Soviet Union) are grate-

fully acknowledged. The research was conducted at the Urgench

State University in Khorezm within the framework of ZEF/

UNESCO landscape restructuring project. A. Khamzina thanks

Prof. Eric Craswell, Prof. Ken Giller, Dr. Antoine Galiana and

Dr. Yves Prin for their advice during the preparation of this

manuscript.

References

Ayaz, F.A., A. Kadioglu and A. Dogru. 1999. Soluble sugar

composition of Elaeagnus angustifolia L. var. orientalia (L.)

Kuntze (Russian olive) fruits. Trop. J. Bot. 23:349–354.

Boddey, R.M., M.B. Peoples, B. Palmer and P.J. Dart. 2000.

Use of the 15N natural abundance technique to quantify

biological nitrogen fixation by woody perennials. Nutr. Cycl.

Agroecosyst. 57:235–270.

Danso, S.K.A., G.D. Bowen and N. Sanginga. 1992. Biological

nitrogen fixation in trees in agroecosystems. Plant Soil

141:177–196.

Dawson, J.O. 1986. Actinorhizal plants: their use in forestry and

agriculture. Outlook Agric. 15:202–208.

Domenach, A.-M., F. Kurdali and R. Bardin. 1989. Estimation

of symbiotic dinitrogen fixation in alder forest by the method

based on natural 15N abundance. Plant Soil 118:51–59.

Dommergues, Y.R. 1987. The role of biological nitrogen fixation

in agroforestry. In Agroforestry a Decade of Development.

Eds. H.A. Steppler and P.K.R. Nair. ICRAF (International

Council for Research in Agroforestry), Nairobi, pp 245–271.





ownloaded from


Dommergues, Y.R. 1995. Nitrogen fixation by trees in relation

to soil nitrogen economy. Fertil. Res. 42:215–230.

Doughton, J.A., I. Vallis and P.G. Saffigna. 1992. An indirect

method for estimating 15N isotope fractionation during

nitrogen fixation by a legume under field conditions. Plant

Soil 144:23–29.

Durinina, E.P. and V.S. Egorov. 1998. Agrohimicheskiy analiz

pochv, rasteniy i udobreniy. MGU, Moscow, 113 p.

Forestry Compendium. 2000. Global module (database). CAB

International, Wallingford, UK.

Friedrich, J.M. and J.O. Dawson. 1984. Soil nitrogen concen-

tration and Juglans nigra growth in mixed plots with nitrogen-

fixing Alnus, Elaeagnus, Lespedeza, and Robinia species. Can.

J. For. Res. 14:864–868.

Galiana, A., J.P. Bouillet and F. Ganry. 2004. The importance

of biological nitrogen fixation by trees in agroforestry. In

Symbiotic Nitrogen Fixation: Prospect for Enhanced Appli-

cation in Tropical Agriculture. Ed. S. Rachid. Science

Publishers, Montpellier, pp 185–199.

Gardner, I.C. 1986. Mycorrhizae of actinorhizal plants.

MIRCEN J. 2:147–160.

Gehring, C., P.L.G. Vlek, L.A.G. de Souza and M. Denich.

2005. Biological nitrogen fixation in secondary regrowth and

primary forests of Central Amazonia. Agric. Ecosyst. Envi-

ron. 111:237–252.

Guo, L.B. and R.M. Gifford. 2002. Soil carbon stocks and land

use change: a meta analysis. Glob. Change Biol. 8:345–360.

Hayes, B. 1976. Planting the Elaeagnus Russian and autumn

olive for nectar. Am. Bee J. 116:74–82.

Heinze, M. and H. Liebmann. 1998. Afforestation of potassium

mine spoil in the southern Harz region. AFZ/Der Wald, Allg.

Forst Z. Wald. Umwelt. 53:1287–1289.

Kachinsky, N.A. 1965. Soil physics. High School Publisher,

Moscow, 322 p.

Katz, G.L. and P.B. Shafroth. 2003. Biology, ecology, and

management of Elaeagnus angustifolia L. (Russian olive) in

western North America. Wetlands 23:763–777.

Khamzina, A., J.P.A. Lamers, M. Worbes, E. Botman and

P.L.G. Vlek. 2006. Assessing the potential of trees for

afforestation of degraded landscapes in the Aral Sea Basin

of Uzbekistan. Agrofor. Syst. 66:129–141.

Khamzina, A., J.P.A. Lamers and P.L.G. Vlek. 2008. Tree

establishment under deficit irrigation on degraded agricultural

land in the lower Amu Darya River region, Aral Sea Basin.

For. Ecol. Manag. 255:168–178.

Khamzina, A., R. Sommer, J.P.A. Lamers and P.L.G. Vlek.

2009. Transpiration and early growth of tree plantations

established on degraded cropland over shallow saline

groundwater table in northwest Uzbekistan. Agric. For.

Meteorol. (in press).

Klute, A. 1986. Water retention: laboratory methods. In

Methods of Soil Analysis. Ed. A. Klute. ASA and SSSA,

Madison, Wisconsin, pp 635–660.

Lamers, J.P.A. and A. Khamzina. 2008. Fuelwood production

in the degraded agricultural areas of the Aral Sea Basin,

Uzbekistan. Bois For. Trop. 297:47–57.

Lamers, J.P.A., I. Bobojonov, A. Khamzina and J. Franz. 2008.

Financial analysis of small-scale forests in the Amu Darya

lowlands of rural Uzbekistan. For. Trees Livelihoods 18:

373–386.

Marcar, N.E. and D.F. Crawford. 2004. Trees for saline

landscapes. Rural Industries Research and Development

Corporation (RIRDC), Canberra, 245 p.

Martius, C., J.P.A. Lamers, P. Wehrheim, A. Schoeller-

Schletter, R. Eshchanov, A. Tupitsa, A. Khamzina, A.

Akramkhanov and P.L.G. Vlek. 2004. Developing sustainable

land and water management for the Aral Sea Basin through

an interdisciplinary research. In Water in Agriculture. Eds. V.

Seng, E. Craswell, S. Fukai and K. Fischer. Australian Center

for International Agricultural Research (ACIAR), Canberra,

ACIAR Proc. No. 116, pp 45–60.

McNiel, R.E. and P.L. Carpenter. 1974. Nitrogen fixation by

woody plant species as measured by the acetylene reduction

assay. HortScience 9:381–382.

Paschke, M.W. 1997. Actinorhizal plants in rangelands of the

western United States. J. Range Manag. 50:62–72.

Peoples, M.B., R.M. Boddey and D.F. Herridge. 2002.

Quantification of nitrogen fixation. In Nitrogen Fixation

at the Millennium. Ed. G.J. Leigh. Elsevier Science,

Amsterdam, pp 357–389.

Ponder, F. 1988. Weed control and autumn-olive affect early

growth and survival of black walnut in a hardwood clearcut.

New For. 2:195–201.

Saigal, S. 2003. Uzbekistan: issues and approaches to combat

desertification. ADB (AsianDevelopment Bank), Tashkent, 51 p.

Shearer, G. and D.H. Kohl. 1986. N2-fixation in field settings:

estimations based on natural 15N abundance. Aust. J. Plant

Physiol. 13:699–756.

Shirokova, Y.I., N. Sharafutdinova and I. Forkutsa. 2000. Use

of electrical conductivity instead of soluble salts for soil

salinity monitoring in Central Asia. Irrig. Drain. Syst. 14:

199–205.

Tjepkema, J.D., C.R. Schwintzer, R.H. Burris, G.V. Johnson

and W.B. Silvester. 2000. Natural abundance of 15N in

actinorhizal plants and nodules. Plant Soil 219:285–289.

Tyurin, I.V. 1975. Metodi sravnitel’noy otzenki soderjaniya

gumusa v pochve. Moscow, 175 p.

Unkovich, M., D. Herridge, M. Peoples, G. Cadish, R.M.

Boddey, K. Giller, B. Alves and P. Chalk. 2008. Measuring

plant-associated nitrogen fixation in agricultural systems.

ACIAR Monogr. 136, 258 p.

Uresk, D.W. and T. Yamamoto. 1994. Field study of plant

survival as affected by amendments to bentonite spoil. Great

Basin Nat. 54:156–161.

vanKessel, C., R.E. Farrel, J.P. Roskoski andK.M.Keane. 1994.

Recycling of the naturally-occurring 15N in an established stand

of Leucaena leucocephala. Soil Biol. Biochem. 26:757–762.

VSN. 1985. Soviet Union Standard Specifications. Pochvennie

iziskaniya dlya meliorativnogo stroitel’stva. In Pochvenno-

meliorativnoe Obosnovanie Proektov Meliorativnogo Stroi-

tel’stva. Moscow, 185 p.

World Bank for Reconstruction and Development. 2002. Global

condition of environment. Tashkent.

Yoneyama, N., N. Yamada, H. Kojima and J. Yazaki. 1984.

Variations of natural 15N abundance in leguminous plants

and nodule fractions. Plant Cell Physiol. 25:1561–1565.


TREE PHYSIOLOGY



ownloaded from


Documents

Nitrogen fixation by Elaeagnus angustifolia in the reclamation of degraded croplands of Central Asia