Phosphorus addition changes belowground biomass and C:N:P stoichiometry of two desert steppe plants under simulated N deposition.

Many studies have reported that increasing atmospheric nitrogen (N) deposition broadens N:phosphorus (P) in both soils and plant leaves and potentially intensifies P limitation for plants. However, few studies have tested whether P addition alleviates N-induced P limitation for plant belowground growth. It is also less known how changed N:P in soils and leaves affect plant belowground stoichiometry, which is significant for maintaining key belowground ecological processes. We conducted a multi-level N:P supply experiment (varied P levels combined with constant N amount) for Glycyrrhiza uralensis (a N fixing species) and Pennisetum centrasiaticum (a grass) from a desert steppe in Northwest China during 2011-2013. Results showed that increasing P addition increased the belowground biomass and P concentrations of both species, resulting in the decreases in belowground carbon (C):P and N:P. These results indicate that P inputs alleviated N-induced P limitation and hence stimulated belowground growth. Belowground C:N:P stoichiometry of both species, especially P. centrasiaticum, tightly linked to soil and green leaf C:N:P stoichiometry. Thus, the decoupling of C:N:P ratios in both soils and leaves under a changing climate could directly alter plant belowground stoichiometry, which will in turn have important feedbacks to primary productivity and C sequestration.

Introduction

It is well known that anthropogenic activities, such as fossil fuel combustion, fertilizer use and intensive animal husbandry have produced a large number of n class="Chemical">nitrogen-containpan>inpan>g compounpan>ds, which have resulted inpan> inpan>creasinpan>g atmospheric pan> class="Chemical">nitrogen (N) deposition[1]. It has been estimated that the global production of reactive N increased from 1.50 × 1013 g N year−1 in 1860 to 1.65 × 1014 g N year−1 in 2000[2]. In China, N deposition amounts have increased at the average rate of 0.041 g m−2 year−1 during 1980–2010[3] and are expected to continue over the next several decades[4]. The ongoing increase in N deposition has accelerated N cycling in terrestrial ecosystems and also altered phosphorus (P) availability to plants[5-7]. Phosphorous has been demonstrated to limit plant growth in most terrestrial ecosystems, either individually or in combination with N[8]. Thus, the increasing P limitation would affect primary production and P cycling in P-limited ecosystems. This raises two fundamental questions under N deposition: (1) how P addition affects P uptake and consequently C:N:P stoichiometry in plants and (2) whether or not P addition alleviates N-induced P limitation for plant growth.

Plant belowground and aboveground processes in plant-soil systems are tightly interlinked and their associations greatly influence ecosystem functions[9,10]. At an individual level, plant belowground organs uptake nutrients from soils and hence link plants and soils together[11]. Changes of soil nutrient availability will directly affect nutrient absorption and assimilation in plant belowground organs, thereby influencing C, N and P relationships in plant leaves[12]. Therefore, to better understand the effects of P addition on plant C:N:P stoichiometry under N deposition, it is necessary to consider the changes in soil properties and leaf eco-physiological characteristics while also taking into account the nutrient processes in plant belowground organs[11]. To date, much of the attention regarding nutrient dynamics and their relationships have been focused on soils and plant leaves. However, a clear explanation on the responses of C:N:P stoichiometry in plant belowground organs to P addition under N deposition is still lacking. If plant belowground C:N:P stoichiometry do change greatly, then plant belowground growth will be altered accordingly. Unfortunately, how plant belowground stoichiometry regulates belowground biomass accumulation under P addition combined with N has not been fully explored yet.

C:N:P ecological stoichiometry is a new approach for studying the interaction between plants and soils under changing global climate[13,14]. It is commonly suggested that elemental stoichiometry is inherently steady, which is important for maintaining ecosystem stability. However, the balance of elemental stoichiometry in soils and plant leaves tend to be decoupled by intensifying climate change[15-17]. In some N limited ecosystems, short-term N addition had inconsistent influence on C status but increased soil N and P availability, consequently, promoting leaf N uptake and synergistic absorption of P[18]. With gradual input of N, N:P becomes disproportionate and plant demand for P increases[19]. These nonsynchronous changes of the three elements generally result in increasing N:P but decreasing C:N in soils and leaves[20,21]. The cycles of C, N and P in plant-soil systems are reciprocal interchanged among aboveground and belowground organs of plants and soils[22]. Thus, the decoupling of C:N:P relationships in soils and leaves are supposed to affect the stoichiometric balance in plant belowground organs, which closely associates with key belowground ecological processes[11]. However, this conjecture has not been widely tested in terrestrial ecosystems, especially in desert steppe ecosystems.

The n class="Disease">Ningxia Hui Autonomous Region, China, covers a large area of desert steppe ecosystem, which is characterized by low soil N availability and critical load of N deposition[23]. A recent study estimates that N deposition has exceeded 4.0 g m−2 year−1 inpan> some places of this region durinpan>g 2000–2010[24]. Therefore, the growth anpan>d stoichiometry of planpan>ts growinpan>g there should be affected accordinpan>gly. The observations from our previous field experiment, which was conducted inpan> 2010 inpan> a desert steppe of Ninpan>gxia, showed that both planpan>t growth anpan>d species number decreased while N addition exceeded 10 g m−2 yr−1 [25,26]. We speculated that over 10 g N m−2 yr−1 might widen N:P anpan>d accelerate P limitation for most species. Thus, we designed the present experiment which inpan>volved six N:P treatments durinpan>g 2011–2013. After 3-year of treatments, we anpan>alyzed planpan>t belowgrounpan>d biomass anpan>d stoichiometry anpan>d their relationships with C:N:P ratios inpan> both soils anpan>d green leaves, respectively. Our objectives were to explore, (1) how P addition affects planpan>t belowgrounpan>d stoichiometry unpan>der simulated N deposition, (2) if anpan>d to what extent P addition alleviates N-inpan>duced P limitation for planpan>t belowgrounpan>d growth anpan>d (3) how the decouplinpan>g of C:N:P ratios inpan> both soils anpan>d leaves affect planpan>t belowgrounpan>d stoichiometry. Our results will be helpful to identify the mechanpan>ism controllinpan>g elemental dynpan>amics withinpan> soils anpan>d planpan>ts inpan> desert steppe ecosystems.

Results

Changes in belowground biomass and root/rhizome to shoot ratio (RSR)

Both N:P supply and species had significant effects on belowground biomass, whereas only species had a significant effect on RSR (Table 1, P < 0.01). Specifically, low P addition (high N:P supply) had insignificant effects on belowground biomass of the two species, while high P addition (≥16 g m−2 year−1 for n class="Species">G. uralensis anpan>d ≥8 g m−2 year−1 for P. centrasiaticum, respectively) greatly inpan>creased belowgrounpan>d biomass of both species (Fig. 1). Inpan> contrast, there were no significanpan>t differences inpan> RSR among six N:P supply treatments. Onpan> average, pan> class="Species">G. uralensis had relatively low belowground biomass but high RSR compared with P. centrasiaticum (Fig. 1 and Table 2).

Table 1

Effects of N:P supply treatment and species and their interaction on belowground growth and C:N:P stoichiometry of the two species (Two-Way ANOVA).

Source		d.f.	F	P
Belowground biomass	N:P supply	5	7.657	0.000
	Species	1	15.164	0.000
	N:P supply × Species	5	0.153	0.977
Root/Rhizome to shoot ratio	N:P supply	5	0.739	0.602
	Species	1	7.974	0.009
	N:P supply × Species	5	0.736	0.604
Belowground C concentration	N:P supply	5	1.037	0.419
	Species	1	54.333	0.000
	N:P supply × Species	5	0.758	0.589
Belowground N concentration	N:P supply	5	5.995	0.000
	Species	1	38.218	0.000
	N:P supply × Species	5	0.578	0.716
Belowground P concentration	N:P supply	5	66.221	0.001
	Species	1	12.539	0.002
	N:P supply × Species	5	24.242	0.001
Belowground C:N ratio	N:P supply	5	5.433	0.002
	Species	1	60.913	0.000
	N:P supply × Species	5	1.629	0.191
Belowground C:P ratio	N:P supply	5	96.161	0.000
	Species	1	55.963	0.000
	N:P supply × Species	5	20.602	0.000
Belowground N:P ratio	N:P supply	5	40.200	0.000
	Species	1	17.994	0.000
	N:P supply × Species	5	3.900	0.000

Figure 1

Effects of N:P supply treatments on belowground biomass and root/rhizome to shoot ratio of the two species. Lowercases above black bars and uppercases above grey bars represent significant differences (P < 0.05) among the N:P treatments for G. uralensis and for P. centrasiaticum, respectively.

Table 2

F values of the differences between the two species under each N:P supply treatment (Independent-Samples T test).

Indices	N10P1	N10P2	N10P4	N10P8	N10P16	N10P32
Belowground biomass	3.161	0.026	10.827**	5.338*	0.470	0.003
Root/rhizome to shoot ratio	3.293	0.021	0.312	1.431	0.117	1.594
Belowground C concentration	3.106	0.431	3.072	0.000	6.823*	0.520
Belowground N concentration	6.258*	2.324	2.183	1.547	0.020	2.968
Belowground P concentration	0.522	0.000	3.798	0.021	2.218	10.077**
Belowground C:N ratio	0.705	1.661	1.641	0.001	2.758	10.178
Belowground C:P ratio	3.707	3.018	4.062	1.182	0.065	5.991*
Belowground N:P ratio	2.138	5.623*	3.821	3.319	0.001	0.486

*and ** represent effects are significant at the 0.01 and 0.05 levels, respectively.

Effects of N:pan class="Chemical">P supply treatment and species and their interaction on belowground growth and C:N:P stoichiometry of the two species (Two-Way ANOVA).

Effects of N:P supply treatments on belowground biomass and root/rhizome to shoot ratio of the two species. Lowercases above black bars and uppercases above grey bars represent significant differences (P < 0.05) among the N:P treatments for n class="Species">G. uralensis anpan>d for P. centrasiaticum, respectively.

F values of the differences betweepan class="Chemical">n the two species under each N:P supply treatment (Independent-Samples T test).

*and ** represepan class="Chemical">nt effects are significant at the 0.01 and 0.05 levels, respectively.

Changes in C:N:P stoichiometry of belowground organs

Both N:P supply and species had large effects on belowground C:N:P stoichiometry (except belowground C), whereas their interaction only had significant effects on belowground P, C:P and N:P (Table 1, P < 0.01). Across the six N:P treatments, there was lack of significant trends in belowground C, N and C:N. In contrast, the belowground P of both species substantially increased with increasing P amount, consequently resulting in decreases of belowground C:P and N:P (Fig. 2). On average, n class="Species">G. uralensis had higher belowgrounpan>d C, C:N anpan>d C:P but lower N, P anpan>d N:P thanpan> P. centrasiaticum (Fig. 2 anpan>d Table 2).

Figure 2

Effects of N:P supply treatments on belowground C:N:P stoichiometry of the two species. Lowercases above black bars and uppercases above grey bars represent significant differences (P < 0.05) among the N:P treatments for G. uralensis and for P. centrasiaticum, respectively.

Effects of N:P supply treatments on belowground C:N:P stoichiometry of the two species. Lowercases above black bars and uppercases above grey bars represent significant differences (P < 0.05) among the N:P treatments for n class="Species">G. uralensis anpan>d for P. centrasiaticum, respectively.

Changes in C:N:P stoichiometry in soils and green leaves

Across the six N:P treatments, the C:P and N:P in soils and green leaves of both species generally decreased with increasing P amount. However, there was no clear effects on C:N both in soils and in green leaf of n class="Species">G. uralensis (Fig. 3). Inpan> general, the averages of C:N, C:P anpan>d N:P inpan> green leaves of pan> class="Species">G. uralensis were higher than those in P. centrasiaticum.

Figure 3

Effects of N:P supply treatments on C:N:P stoichiometry in both soils and green leaves of the two species. Lowercases above black bars and uppercases above grey bars represent significant differences (P < 0.05) among the N:P treatments for G. uralensis and for P. centrasiaticum, respectively.

Effects of N:P supply treatments on C:N:P stoichiometry in both soils and green leaves of the two species. Lowercases above black bars and uppercases above grey bars represent significant differences (P < 0.05) among the N:P treatments for n class="Species">G. uralensis anpan>d for P. centrasiaticum, respectively.

Relationships between plant belowground biomass and C:N:P stoichiometry

Belowground biomass of papan class="Chemical">n class="Species">G. uralensis was highly related to belowgrounpan>d C, belowground N:P, soil C:P, soil N:P and also green leaf C:P, respectively. For P. centrasiaticum, belowground biomass was only significantly related to belowground C, soil N:P, green leaf C:P and also green leaf N:P, respectively (Fig. 4).

Figure 4

Relationships between belowground biomass and C:N:P stoichiometry. Black circles fitted with solid lines are for G. uralensis (R12), white circles fitted with dot lines are for P. centrasiaticum (R22), respectively. (a–f), (g–i) and (j–l) represent for the relationships between belowground biomass and belowground C:N:P, between belowground biomass and soil C:N:P and between belowground biomass and green leaf C:N:P, respectively.

Relationships between belowground biomass and C:N:P stoichiometry. Black circles fitted with solid lines are for n class="Species">G. uralensis (R12), white circles fitted with dot linpan>es are for P. centrasiaticum (R22), respectively. (a–f), (g–i) anpan>d (j–l) represent for the relationships between belowgrounclass="Chemical">pan>d biomass and belowgrounpan>d C:N:P, between belowgrounpan>d biomass anpan>d soil C:N:P anpan>d between belowgrounpan>d biomass anpan>d green leaf C:N:P, respectively.

Relationships between plant belowground C:N:P stoichiometry and soil C:N:P stoichiometry

Positive lipan class="Chemical">near relationships were found between belowground N:P and soil C:P and also between belowground N:P and soil N:P for n class="Species">G. uralensis, respectively (Fig. 5). Strong positive relationships were present between belowground C:P and soil C:P, between belowground C:P and soil N:P, between belowground N:P and soil C:P, as well as between belowground N:P and soil N:P for P. centrasiaticum, respectively.

Figure 5

Relationships between belowground C:N:P stoichiometry and soil C:N:P stoichiometry. Black circles fitted with solid lines are for G. uralensis (R12), white circles fitted with dot lines are for P. centrasiaticum (R22), respectively. (a–f), (g–l) and (m–r) represent for the relationships between belowground C:N:P and soil C:N, between belowground C:N:P and soil C:P and between belowground C:N:P and soil N:P, respectively.

Relationships between belowground C:N:P stoichiometry and soil C:N:P stoichiometry. Black circles fitted with solid lines are for n class="Species">G. uralensis (R12), white circles fitted with dot linpan>es are for P. centrasiaticum (R22), respectively. (a–f), (g–l) anpan>d (m–r) represent for the relationships between belowgrounpan>d C:N:P anpan>d soil C:N, between belowgrounpan>d C:N:P anpan>d soil C:P anpan>d between belowgrounpan>d C:N:P anpan>d soil N:P, respectively.

In contrast, a negative linear relationship was present between belowground C and soil N:P for n class="Species">G. uralensis. The strong negative linpan>ear relationships were present between belowgrounpan>d P anpan>d soil C:P anpan>d also between belowgrounclass="Chemical">pan>d P and soil N:P for P. centrasiaticum, respectively.

Relationships between plant belowground C:N:P stoichiometry and green leaf C:N:P stoichiometry

Positive linear relationships were found between belowground C:P and green leaf C:P, between belowground C:P and green leaf N:P, between belowground N:P and green leaf C:P and also between belowground N:P and green leaf N:P for n class="Species">G. uralensis, respectively (Fig. 6). Againpan>, positive linpan>ear relationships were observed between belowgrounpan>d P anpan>d green leaf C:N, between belowgrounpan>d C:P anpan>d green leaf C:P, between belowgrounpan>d C:P anpan>d green leaf N:P, between belowgrounpan>d N:P anpan>d green leaf C:P, as well as between belowgrounpan>d N:P anpan>d green leaf N:P for P. centrasiaticum, respectively.

Figure 6

Relationships between belowground C:N:P stoichiometry and green leaf C:N:P stoichiometry. Black circles fitted with solid lines are for G. uralensis (R12), white circles fitted with dot lines are for P. centrasiaticum (R22), respectively. (a–f), (g–l) and (m–r) represent for the relationships between belowground C:N:P and green leaf C:N, between belowground C:N:P and green leaf C:P and between belowground C:N:P and green leaf N:P, respectively.

Relationships between belowground C:N:P stoichiometry and green leaf C:N:P stoichiometry. Black circles fitted with solid lines are for n class="Species">G. uralensis (R12), white circles fitted with dot linpan>es are for P. centrasiaticum (R22), respectively. (a–f), (g–l) anpan>d (m–r) represent for the relationships between belowgrounpan>d C:N:P anpan>d green leaf C:N, between belowgrounpan>d C:N:P anpan>d green leaf C:P anpan>d between belowgrounpan>d C:N:P anpan>d green leaf N:P, respectively.

Negative linear relationships were found between belowground P and green leaf C:P and also between belowground P and green leaf N:P for n class="Species">G. uralensis. Negative relationships were also observed between belowgrounpan>d P anpan>d green leaf C:P, between belowgrounpan>d P anpan>d green leaf N:P, between belowgrounpan>d C:P anpan>d green leaf C:N, as well as between belowgrounpan>d N:P anpan>d green leaf C:N for P. centrasiaticum, respectively.

Discussion

Previous research has suggested that an wider N:P supply (i.e. increasing N addition) can promote the growth of plants limited by N, whereas a reduced N:P supply (i.e. increasing P addition) is expected to increase the growth of plants limited by P[27]. We observed that low P addition (high experimental N:P supply) had little effects on the belowground biomass of both species. With the increase of P amounts (the thresholds were 16 g m−2 year−1 for n class="Species">G. uralensis anpan>d 8 g m−2 year−1 for P. centrasiaticum, respectively), the belowgrounpan>d biomass of both species gradually inpan>creased as well (Fig. 1), which is similar to the observations made from a grass species pan> class="Species">Setaria sphacelata[10]. One possible reason would be that the tested soil is poor in P availability, which was exacerbated by N addition. More importantly, the sampled soil is slightly alkaline, in which P is usually bound to carbonates[28]. Thus, low P addition might not be enough to overcome the carbonate barrier, rendering P unavailable to plants[29]. With increasing P addition, soil P availability increased and N-induced P pressures on plants were alleviated[5], resulting in the increases of belowground growth. Our results indicate that moderate P addition is of great benefit to alleviate N-induced P limitation for the belowground growth of desert steppe plants. However, high P addition (exceeding 8 g P m−2 year−1) would cause P toxicity and N limitation[30], which possibly slowed down the rate of belowground biomass accumulation of P. centrasiaticum (Fig. 1).

Plant biomass allocations between aboveground and belowground organs provide key information for connecting aboveground productivity and belowground C sequestration[31]. It reflects the balance between aboveground resources (light and n class="Chemical">CO2) anpan>d belowgrounpan>d resources (pan> class="Chemical">water and nutrients)[32] and is highly plastic in response to environmental changes[33,34]. In nutrient poor ecosystems, plant growth will shift from nutrient limitation to light limitation under increasing nutrient enrichment[35]. To grow faster, plants often allocate more biomass to aboveground organs for achieving high leaf photosynthesis and thus producing more dry matter[27]. Some studies found that N addition mainly facilitated aboveground growth[36] while P addition might be more beneficial for the biomass accumulation in belowground organs[37]. In the present study, the aboveground and belowground biomass of both species showed similar responses to increasing P addition under simulated N deposition as observed in other study[5], thereby resulting in the insignificant differences of RSR among the treatments (Fig. 1). Compared with the previous studies involved with the single application of N or P[36,37], our results may suggest that plant biomass allocation becomes more complicated under the combined application of N and P[34].

Numerous studies have concluded that C:N:P ratios both in soils and in plant leaves would decouple under increasing nutrient enrichment[5,16,17,20]. Due to the differences in nutrient storage and metabolism function, plant belowground organs are considered to be less sensitive to environmental changes than plant leaves. Therefore, belowground C:N:P stoichiometry of both species were assumed to be less affected by P addition in the present study. However, we observed clear changes of C:P and N:P either in soils and green leaves or in belowground organs of both species across the treatments, which are inconsistent with the insignificant response of a grass species n class="Species">S. sphacelata[10]. Onpan>e possible explanpan>ation would be that inpan>creasinpan>g P supply improved soil P availability anpan>d thus promoted the P uptake of both species as reported by anpan>other study[38]. This more positive response of belowgrounpan>d P thanpan> C anpan>d N consequently caused chanpan>ges inpan> belowgrounpan>d stoichiometric relationships. The inpan>creased belowgrounpan>d P uptake also reflected the alleviated P limitation for both species. Our results may support the finpan>dinpan>gs that stoichiometric balanpan>ce inpan> belowgrounpan>d organpan>s canpan> also be destroyed by inpan>tensified environmental chanpan>ges[36]. The altered belowgrounpan>d C:N:P relationships would inpan> turnpan> have importanpan>t feedbacks to planpan>t productivity anpan>d C sequestration through its inpan>fluences on nutrient absorption, root turnpan>over, microbial activity anpan>d other belowgrounpan>d ecological processes[36].

Grasses are generally thought to possess lower N and P concentrations and be more easily affected by N and P fertilization than N-fixing species[8,39]. In the present study, we found that P. centrasiaticum had unexpectedly high belowground N (21.14 mg g−1 vs. 16.59 mg g−1) and P (2.47 mg g−1 vs. 2.24 mg g−1) concentrations compared with n class="Species">G. uralensis. Onpan> the one hanpan>d, the 10 g N m−2 yr−1 might have more positive effect on belowgrounpan>d N uptake of P. centrasiaticum, which is similar with the observations from previous study[39]. Accordinpan>gly, the belowgrounpan>d organpan>s of P. centrasiaticum needed to absorb more P to balanpan>ce the C:N:P relationship as reported inpan> previous studies[18,40]. Inpan> this case, inpan>creased N availability would simultanpan>eously stimulate the activity of P-minpan>eralizinpan>g enzymes both inpan> soils anpan>d on planpan>t belowgrounpan>d organpan>s[41] anclass="Chemical">pan>d also increase the root/rhizome colonization of P-acquirinpan>g arbuscular mycorrhizal funpan>gi[42]. Both of those approaches could provide more available P for belowgrounpan>d uptake anpan>d assimilation of P. centrasiaticum. Onpan> the other hanpan>d, the belowgrounpan>d growth of P. centrasiaticum was more greatly stimulated by inpan>crease of P addition, which promoted it uptake more N anpan>d P from soils. Inpan> summary, the greater positive response of N uptake to N anpan>d P fertilization likely resulted inpan> higher N:P inpan> belowgrounpan>d organpan>s of P. centrasiaticum thanpan> that of pan> class="Species">G. uralensis (Fig. 2 and Table 2).

Previous studies have reported that C:N:P stoichiometry in soils and plants are tightly linked[9,11]. Plant C:N:P stoichiometry reflects litter decomposition quality and thus directly determines soil N and P availabilities, while soil C:N:P stoichiometry regulates microbial activity and also plant N and P uptakes. In the present study, we found C:P and N:P in belowground organs of the two species were highly related to C:P and N:P both in soils and in green leaves as reported in other studies[9,12,43], especially P. centrasiaticum. The roots of N-fixing species can build symbiotic relationships with nodule bacteria, which in turn aid N-fixing species obtain N more easily and therefore depend less on soil N than other growth forms. Thus, the self-adjusting N strategy of n class="Species">G. uralensis probably contributed to less susceptibility of belowgrounpan>d C:N:P stoichiometry to chanpan>ginpan>g C:N:P relationships inpan> soils when compared with P. centrasiaticum. Our results suggest that the decouplinpan>g of C:N:P stoichiometry inpan> soils anpan>d leaves unpan>der global climate chanpan>ge[15,17] will directly affect the elemental stoichiometry balanpan>ce inpan> planpan>t belowgrounpan>d organpan>s, which will inpan> turnpan> have importanpan>t inpan>fluences on leaf nutrient uptake, soil nutrient availability anpan>d also nutrient cyclinpan>g inpan> planpan>t-soil systems.

The C:N:P stoichiometry in both soils and plants could indicate nutrient limitation patterns and also relate to nutrient use efficiency[44,45]. Therefore, the changes of C:N:P relationships in soils and plants under changing environments would directly affect plant growth and biomass accumulation. In the present study, we found that the belowground biomass of n class="Species">G. uralensis inpan>creased with decreasinpan>g belowgrounpan>d N:P, soil C:P, soil N:P anpan>d also green leaf C:P, while the belowgrounpan>d biomass of P. centrasiaticum improved with reducinpan>g soil N:P, green leaf C:P anpan>d also green leaf N:P, respectively. Sinpan>ce high C:P anpan>d N:P have been considered as more P limited thanpan> N[27], these negative relationships may further prove that inpan>creasinpan>g P addition alleviates soil P limitation. Thus, biomass accumulation is promoted inpan> belowgrounpan>d organpan>s of both species through its negative inpan>fluences on soil C:P anpan>d N:P. The results broadly suggest that the elemental stoichiometry both inpan> soils anclass="Chemical">pan>d in planpan>ts play importanpan>t roles inpan> drivinpan>g planpan>t biomass production. However, more field experiments are needed to test whether this statement holds inpan> natural communpan>ities.

Conclusion

Numerous studies have verified that chronic N deposition induces the imbalance of N:P and the increase of P limitations in grasslands[46,47]. Although the monitored N deposition amount in desert steppe is lower than the data reported in other ecosystems[3], low soil N availability and a critical load of N deposition may result in high sensitivity to chronic and low-level N deposition in a desert steppe. Our results showed that increasing P addition enhanced belowground P of both species and thus altered belowground C:N:P stoichiometry, which was tightly linked to soil and leaf C:N:P stoichiometry. Moderate P addition regulated the balance between soil P supply and plant P demand, consequently, alleviating N-induced P limitations for plant growth. The self-adjusting N strategy of n class="Species">G. uralensis might make it keep high inpan>herent stability of C:N:P stoichiometry despite of great stoichiometric chanpan>ge inpan> soils. Our results canpan> provide scientific basis for adaptive manpan>agement of pan> class="Disease">fragile ecosystems under global climate change. However, the present study was based on a pot-cultured experiment and only performed for three years. Long-term simulated field experiments that test the comparisons among multiple growth forms are urgently needed to further improve our understanding of how plant belowground stoichiometry responds to altering environments and of the adaptation of fragile ecosystems to global climate change.

Materials and Methods

Study site

This study was conducted at Sidunzi Grassland Research Station, Yanchi County (37°64′N, 106°50′E, 1450 m a.s.l), n class="Disease">Ningxia Hui Autonomous Region, Chinpan>a. This region lies on the southwest edge of the Mu Us Sanpan>dy Lanpan>d anpan>d has a typical continental climate inpan> the moderate temperate zone. Meanpan> anpan>nual precipitation is 289 mm, with over 70% of it occurrinpan>g inpan> the growinpan>g season (May to September). Meanpan> anpan>nual evaporation is 2132 mm. Meanpan> anpan>nual temperature is 7.7 °C anpan>d meanpan> monthly temperatures ranpan>ge from −8.9 °C inpan> Janpan>uary anpan>d 22.5 °C inpan> July. The major soil is classified as Aridisol (FAO classification). The study site has been fenced from domestic anpan>imals sinpan>ce 2001. The planpan>t communpan>ity is chiefly composed of Lespedeza potanpan>imill, P. centrasiaticum, pan> class="Species">G. uralensis, Sophora alopecuroides, Cleistogenes squarrosa, Agropyron cristatum, Stipa capillata, Oxytropis aciphylla, Caragana intermedia, Cynanchum komarovii and other grasses and shrubs.

Experimental design

The selected species were n class="Species">G. uralensis and P. centrasiaticum, both of which are typical perennial species for livestock pasture, winpan>d prevention anpan>d sanpan>d fixation inpan> northwesternpan> Chinpan>a. pan> class="Species">G. uralensis is an N-fixing species and its roots can build symbiotic relationships with nodule bacteria. P. centrasiaticum is a grass and is characterized by functional cross rhizomes. Due to their important economic and ecological value, the plantation of G. uralensis and P. centrasiaticum have been developed since the early 21th century.

In early April 2011, 36 n class="Chemical">polyvinyl chloride pots (each 160 mm in diameter and 500 mm inpan> height) were vertically buried inpan>to soils with about 50 mm remainpan>inpan>g above grounpan>d level. Each pot was filled with about 10.0 kg of soils collected from the native habitats of the two studied species. The basic soil properties were measured before tranpan>splanpan>tation. The organpan>ic C, total N, total P, pan> class="Chemical">NH4+-N, NO3−-N, available P and pH were 1.89 g kg−1, 0.22 g kg−1, 0.31 g kg−1, 1.06 mg kg−1, 8.26 mg kg−1, 13.16 mg kg−1 and 8.53, respectively. G. uralensis and P. centrasiaticum were planted together in each pot and two seedlings were selected for each species. All seedlings of each species were similar in morphology, specifically 10–12 cm in height and 0.4–0.5 cm in basal diameter and 5–6 in leaf number for G. uralensis while 8–10 cm in height and 0.3–0.4 cm in basal diameter and 4–5 in leaf number for P. centrasiaticum, respectively. After two weeks, the healthier individual of each species was kept in each pot.

To determine to what extent P addition might alleviate N-induced P limitations for plant growth, six N:P levels were designed and three replicates were chosen for each level. The six levels were treated with the same N amount (10 g N m−2 year−1) but with different P amounts: 1, 2, 4, 8, 16 and 32 g P m−2 year−1, thus producing six experimental N:P supply treatments (N10P1, N10P2, N10P4, N10P8, N10P16 and N10P32, respectively). N and P fertilizers were uniformly applied to each pot from May to August during 2011–2013. In order to increase the utility of both fertilizers and avoid high P poison situations, n class="Chemical">NH4NO3 anpan>d pan> class="Chemical">KH2PO4 were dissolved into water and were added 2–4 times per week. All pots were placed in a flat field and thus received the same precipitation, temperature, light and other environmental factors.

Field sampling

In late August of 2013, the belowgroupan class="Chemical">nd organs of each species were harvested (roots for n class="Species">G. uralensis and rhizomes for P. centrasiaticum, respectively). Thirty fully expanclass="Chemical">pan>ded green leaves and the rest of the aboveground organs of each species were also collected. Soils were randomly sampled with an auger at a depth of 0–10 cm. All samples were stored inpan> an inpan>sulated can and immediately taken to the laboratory for further analysis.

Laboratory methods and calculations

Belowground organs of each species were rinsed with distilled n class="Chemical">water anpan>d dried inpan> anclass="Chemical">pan> oven set to 75 °C for 48 hours to measure belowgrounpan>d biomass per pot. Green leaves anpan>d the rest of abovegrounpan>d organpan>s were dried at 65 °C for 48 hours to measure abovegrounpan>d biomass per pot. RSR of each species was calculated with the ratio between belowgrounpan>d biomass anpan>d abovegrounpan>d biomass. Total C, N anpan>d P concentrations were anpan>alyzed after samples were finpan>ely grounpan>d inpan> a Wiley Mill anpan>d passed through a 40-mesh sieve. Total C was measured usinpan>g the pan> class="Chemical">K2MnO4 volume method, while total N was analyzed with the Kjeldahl acid-digestion method using an Alpkem AutoAnalyzer (Kjektec System 2300 Distilling Unit, Sweden) and total P was determined colorimetrically at 700 nm after reaction with a molybdenum-antimony solution, respectively.

Soil samples were divided into two parts with the first part analyzed for organic C (n class="Chemical">K2MnO4 volume method), total N (Kjeldahl acid-digestion method) anpan>d total P (class="Chemical">pan> class="Chemical">HCLO4-H2SO4 method) after air drying. The second part was for the immediate analyses of NH4+-N and NO3−-N (both Discontinuous Flow Analyzer, Germany) and available P (NaHCO3 method).

Statistical analysis

SPSS version 13.0 (SPSS Inc., Chicago, IL, USA) was used for statistical analyses. The K-S test was used to confirm normality. The two-way ANOVA was performed to test the effects of N:P supply treatments and species and their interaction on belowground biomass and C:N:P stoichiometry. The one-way ANOVA was performed to test the effects of N:P supply treatments on aboveground biomass, belowground biomass, RSR, belowground C:N:P stoichiometry, green leaf C:N:P stoichiometry and also soil C:N:P stoichiometry, respectively. The independent T test was performed to determine the differences in belowground biomass and C:N:P stoichiometry between the two species under each N:P supply treatment. Linear regression analysis was used to describe the relationships between belowground C:N:P stoichiometry and biomass, between belowground C:N:P stoichiometry and soil C:N:P stoichiometry and also between belowground C:N:P stoichiometry and green leaf C:N:P stoichiometry, respectively. Data are presented as means ± standard error (SE, n = 3).