R E S E A R CH A R T I C L E

Temporal heterogeneity has no effect on the direction ofsuccession in abandoned croplands in a semiarid area ofnorthwest China

An Hu | Sheng-Hua Chang | Xian-Jiang Chen | Fu-Jiang Hou | Zhi-Biao Nan

State Key Laboratory of Grassland Agro-

Ecosystems, Key Laboratory of Grassland

Livestock Industry Innovation, Ministry of

Agriculture, College of Pastoral Agriculture

Science and Technology, Lanzhou University,

Lanzhou, Gansu, PR China

Correspondence

Fu-Jiang Hou, State Key Laboratory of

Grassland Agro-Ecosystems, Key Laboratory of

Grassland Livestock Industry Innovation,

Ministry of Agriculture, College of Pastoral

Agriculture Science and Technology, Lanzhou

University, Lanzhou, 730000, Gansu, PR

China.

Email: cyhoufj@lzu.edu.cn

Funding information

Changjiang Scholar Program of Chinese

Ministry of Education, Grant/Award Number:

IRT17R50; Chinese Academy of Sciences,

Grant/Award Number: XDA2010010203; Key

R & D Program of Ningxia Hui Autonomous

Region, Grant/Award Number:

2019BBF02001; National Natural Science

Foundation of China, Grant/Award Number:

31172249

Abstract

Space for time substitution has been widely used in the study of succession in aban-

doned croplands worldwide but it is often accompanied by time heterogeneity. How-

ever, the effect of temporal heterogeneity on succession dynamics over decades is

not well understood. Here, we studied croplands with the same history that were

abandoned in 1998, 1999, and 2000 in northwest China (temporal heterogeneity)

and continuously monitored for vegetation characteristics for 10 years. Aboveground

biomass, taxonomic, and functional diversity were compared in the three cropland

groups over time. Our results showed that the direction of succession in the three

cropland groups was similar, from a community with single dominant species with

higher aboveground biomass to a community with multiple dominant species with

lower aboveground biomass. Taxonomic and functional diversity increased rapidly in

the first 5–6 years, followed by slow increase, decrease, or stabilization. In years 1–5,

aboveground biomass, taxonomic, and functional diversity were affected by rainfall,

time of abandonment, time since abandonment, or their interactions. However, in

years 6–10, biomass was only affected by rainfall, and functional evenness was

affected by rainfall and time since abandonment; there was no time of abandonment

effect. We conclude that temporal heterogeneity can initially affect the succession

process but has no effect on the direction of community succession in the longer

term. Our findings provide evidence for using space for time substitution with tem-

poral heterogeneity to study succession in abandoned croplands in semiarid areas.

K E YWORD S

abandoned cropland, aboveground biomass, functional diversity, taxonomic diversity,

temporal heterogeneity

1 | INTRODUCTION

Heterogeneity is increasingly recognized as a foundational character-

istic of ecological systems (Collins et al., 2018). In this context, hetero-

geneity refers specifically to spatiotemporal variability of an ecological

factor (McIntosh, 1991). The spatiotemporal heterogeneity can influ-

ence plant community structure and soil physicochemical properties

(Limb, Hovick, Norland, & Volk, 2018; Zhang et al., 2017a). The space

for time substitution approach was used in early ecological studies

(Smith, 1940) as well as recent studies in ecology (Mcgranahan

et al., 2012). Recently, the space for time substitution approach has

again become a topic of interest in the context of community succes-

sion, especially in the succession of abandoned croplands (Thomas,

Knops, & Dave, 2018; Zhang et al., 2018a). Studies of natural and arti-

ficial revegetation of abandoned croplands during succession have

focused on changes in vegetation composition (Hodapp et al., 2018),

Received: 18 January 2020 Revised: 24 August 2020 Accepted: 22 September 2020

DOI: 10.1002/ldr.3780

Land Degrad Dev. 2020;1–10. wileyonlinelibrary.com/journal/ldr © 2020 John Wiley & Sons, Ltd. 1

structure (Zhang, Ren, et al., 2018b), coverage, species diversity (Veen,

Van Der Putten, & Bezemer, 2018), functional diversity (Thomas

et al., 2018), soil properties (Deng et al., 2018), and soil microbes (Zhang

et al., 2017b; Zhong, Yan, Wang, Wang, & Shangguan, 2018). However,

space for time substitution in succession studies in abandoned croplands

is also accompanied by time heterogeneity, the result of cropland rep-

resenting different succession years at T0, which means these croplands

will have been abandoned in different years (Figure 1).

Secondary succession in abandoned cropland without human

interference is an effective means of restoring degraded ecosystems

under extremely poor environmental conditions (Benjamin, Domon, &

Bouchard, 2005; Wang, Liu, & Xu, 2009). The agriculture-pasture eco-

tone in northern China especially in the Loess Plateau gully region is

an excellent example. Characterized by low yet highly variable annual

rainfall, extensive cultivation has been important for farming produc-

tion. The Loess Plateau is now one of the most vulnerable areas in the

F IGURE 1 A conceptual model for the succession of abandonedcroplands in different years. The x-axis (0, t1, t2, and t3) representstime since abandonment. Space for time substitution generally occursat T0 [Colour figure can be viewed at wileyonlinelibrary.com]

F IGURE 2 Taxonomic diversity (left) and functional diversity (right) over time for each abandoned cropland A1998, A1999, and A2000.(a) Species richness, (b) Shannon–Wiener index, (c) species evenness, (d) modified functional attribute diversity (MFAD), (e) functional divergence,and (f) functional evenness. Values represent the mean ± SE for the three cropland groups (replicate succession plots in each cropland). *p < 0.05,**p < 0.01, and ***p < 0.001 is compared among groups of A1998, A1999, and A2000 [Colour figure can be viewed at wileyonlinelibrary.com]

2 HU ET AL.

world owing to poor ecological conditions and pervasive soil erosion.

Vegetation restoration is an effective way to manage water loss and

soil erosion as well as promote restoration of ecosystem services

(Valkó et al., 2016). To solve the problem of ecological degradation on

the Loess Plateau, the Chinese Government initiated a nationwide

ecological programme called the Grain for Green Program to prevent

further soil erosion and improve the quality of the degraded soil,

focusing on restoring vegetation by converting cropland to forest and

grassland. Begun in 1999, this Programme is a major ecological under-

taking; it includes large financial investment, strong policy, wide

involvement, and high levels of public participation in China (Zhang,

Ren, et al., 2018b). The local grassland species on the Loess Plateau

includes herbaceous plants, of which Artemisia capillaris Thunb., Stipa

bungeana Trin., and Lespedeza davurica (Laxm.) Schindl, are dominant

species, accounting for more than 80% of the total biomass (Hu,

Zhang, Chen, Chang, & Hou, 2019).

Rainfall and time since abandonment are the main factors that

affect recovery in abandoned croplands on the Loess Plateau. Rainfall

can directly affect aboveground and underground biomass, and time

of succession has cumulative effects on species diversity and ecosys-

tem productivity (Xu, Xu, & Huang, 2011). Biodiversity and ecosystem

productivity are important indicators for measuring the degree of deg-

radation and restoration in abandoned croplands, and consequently

play important roles in the process of succession.

For this study, we selected croplands in northwest China that

were abandoned in different years (temporal heterogeneity) and con-

ducted 10 years of fixed-point, continuous observation. We analyzed

the plant community structure and six characteristics of diversity for

different times since abandonment. We hypothesize that temporal

heterogeneity may have a different direction during early cropland

succession, but that ultimately, cropland succession proceeds in same

direction (Figure 1). Our study assessed the temporal heterogeneity

by evaluating (a) taxonomic diversity (species richness, Shannon–

Wiener index, and species evenness), and functional diversity (modi-

fied functional attribute diversity, functional divergence, and func-

tional evenness), (b) aboveground biomass and the stability of

biomass, (c) nonmetric multidimensional scaling (NMDS) for biomass

of each aboveground species. In addition, structural equation model-

ing (SEM) was used to model the direct and indirect effects of rainfall

and time since abandonment on total biomass. This study was

designed to assess whether temporal heterogeneity can be used in

succession studies after land is abandoned.

2 | MATERIALS AND METHODS

2.1 | Study site

The study site is in Huanxian County, eastern Gansu Province, north-

west China (37.14�N, 106.84�E) in the Loess Plateau Ravine at an ele-

vation of �1700 m. The region has a mean daily air temperature of

7.1�C. Average annual precipitation for 1998–2010 was 274 mm but

is highly variable (range 149–433 mm; coefficient of variation = 28%;

Figure S1). More than 70% of the total rainfall occurs from July to

September, and the mean annual potential evaporation is 1,993 mm

(Chen, Hou, Matthew, & He, 2017). Spring and autumn are typically

short, summer is hot and humid, and winter is long and cold. The soil

is classified as sandy, free-draining loess and the rangeland is a typical

temperate steppe (Ren et al., 2008).

The main crops grown in the study area prior to abandonment

were potato (Solanum tuberosum L.), maiz (Zea mays L.), foxtail millet

(Setaria italica [L.] P. Beauv.), and buckwheat (Fagopyrum esculentum

Moench) (Xu, 2015). Farmyard manure was applied to crops before

planting but stopped after abandonment. The area has been part of

the Grain for Green Program since 1999, and this has sought to pre-

vent soil erosion and improve the quality of the degraded soil by

focusing on vegetation restoration through conversion of cropland to

forests and grassland. In this study, we selected croplands abandoned

in the years 1998, 1999, and 2000 (hereafter named A1998, A1999,

and A2000, respectively) and monitored aboveground vegetation

characteristics for 10 years. Four abandoned croplands (area from

5,000 to 6,000 m2) were selected each year to provide replication.

The three different years of abandoned croplands created a temporal

heterogeneity. Considering the strong influence of tillage history on

succession (Benjamin et al., 2005), we selected only croplands that

had same farming history, planted with foxtail millet for at least

F IGURE 3 (a) Aboveground biomass (g m−2) in three croplandgroups A1998, A1999, and A2000 over 10 years. Values representthe mean ± SE for the three treatments; *p < 0.05 is compared amonggroups of A1998, A1999, and A2000. (b) Stability of biomassproduction (mean/SD) of the three cropland groups in three timeperiods of 10 years. Error bars represent ±SE [Colour figure can beviewed at wileyonlinelibrary.com]

HU ET AL. 3

5 years before abandonment. The croplands had been cultivated using

a hoe and plough for more than 20 years. After abandonment, there

were no artificial disturbances (grazing, mowing, or fertilization) were

imposed allowing natural succession to commence. At the peak of the

growing season in August of each year, we randomly selected four

1 m2 quadrats within each abandoned cropland. This provided

16 quadrat measurements in each commencement year annually,

480 over the 10-year study period.

2.2 | Aboveground characterization

In each quadrat, plant density, tiller, and branch numbers for graminoid

and non graminoid species, respectively, was measured. Five plants were

randomly selected for each species per quadrat to measure plant height

(absolute plant height), crown breadth (maximum distance from the top

of the branches or tillers), and branches or tillers numbers. In addition, the

entire aboveground biomass from each quadrat was clipped and sorted

into dead and live fractions, and then dried and weighed. The total above-

ground biomass only included the live fractions.

2.3 | Data analysis

Taxonomic and functional diversity were calculated per abandoned

cropland using the average aboveground biomass for each species

and the 10 traits (Table S1).

Community diversity was estimated by measuring plant species

richness (number of plant species), Shannon's diversity index (H),

H= −XS

ipiln pið Þ,

F IGURE 4 A structural equation modeling describing the effects of rainfall and time since abandonment on total biomass (a), andstandardized total effects (STE, the sum of direct and indirect effects from each variable) from the SEM on total biomass (b). Numbers labeling thearrows are the standardized path coefficients, indicating the strength of the relationships. The grey dotted line indicates no significant correlation,the black line indicates a significant positive correlation, and the red line indicates a significant negative correlation. Significance levels of eachpredictor are *p < 0.05, **p < 0.01, and ***p < 0.001 [Colour figure can be viewed at wileyonlinelibrary.com]

4 HU ET AL.

Where: pi is the biomass of i species and accounts for total species

biomass.

The proportion of individuals in community was estimated by the

Pielou index (E),

E =H=lnS,

Where: S is the total number of species in the community.

The functional diversity was estimated using modified functional

attribute diversity (MFAD), which measures the dispersion of species

in the functional traits space.

MFAD=XNh=1

XNk =1

dhk

!=N;

dhk =Xpi=1

ahiakij j=Xpi=1

max ahi,akif g,

Where: dhk is the dissimilarity between functional units h and k, ahi is

the affinity of functional unit h to trait i and aki is the affinity of func-

tional unit k to trait i, and N is the number of functional units

(Schmera, Er}os, & Podani, 2009).

Functional divergence (Fdiv) is the deviance of species from the

mean distance to the center of the multivariate trait space weighted

by the relative cover of the species. Low functional divergence indi-

cates that abundant species have trait values that are close to the

community-weighted trait average, while high functional divergence

indicates that the most abundant species have more extreme trait

values (Botta-Dukát, 2010).

Fdiv =XNh=1

XNk >1

dhkphpk ,

Functional evenness (Feve) quantifies the distribution of species

traits in the occupied trait space (Villéger, Mason, & Mouillot, 2008).

Feve =XS−1

b=1

min PEWb1=ðS−1Þ½ �−1= S−1ð Þg= 1−1=ðS−1Þ½ �( )

EWb =EWbPS−1

b=1EWb

,

EWb = dhk= ph + pkð Þ:

Before calculating functional diversity, plant traits data were stan-

dardized using a Z-transformation to obtain a mean of zero and a

standard deviation of one, because they were measured in different

units.

Succession stages were to be divided across multiple years or

cohorts of years (Zhong et al. 2018). We thus divided the succession

time into three cohorts, that is, 1–4, 4–7, 7–10 years, as this allowed

us to include the full 10 years of experimental data and examine tem-

poral changes in stability of aboveground biomass (Veen et al., 2018).

We also used NMDS to visualize how the composition of the plant

community changed over time. NMDS was carried out in R version

3.5 (R Development Team, 2013).

One-way ANOVA was used to determine the differences in the

aboveground biomass, stability of biomass, taxonomic diversity (spe-

cies richness, Shannon–Wiener index, species evenness), and func-

tional diversity (MFAD, functional divergence, functional evenness)

among three cropland groups at same time since abandonment in suc-

cessional stages. The statistical significance was considered at

p < 0.05. Mean values (±SE) are presented in Figures 3–5.

As data were not normally distributed (Shapiro–Wilk test), a gen-

eralized linear model (GLIMMIX procedure) was applied to quantify

the effects of rainfall (R, from the previous year's September to cur-

rent August), time since abandonment (Y, one to 10 years), and time

of abandonment (T, the year each of the three cropland groups aban-

doned: A1998, A1999, and A2000) on the aboveground biomass, tax-

onomic diversity (species richness, Shannon–Wiener index, and

species evenness), and functional diversity (MFAD, functional diver-

gence, and functional evenness; Table 1). The model is y = R + Y + T

+ R × Y + R × T + T × Y + R × T × Y+ Γ + ε, where Γ is the random

effect of replicate, and ε is the model error. Based on the observation

of the changes in each indicator, we divided the years into two epi-

sodes, that is, 1–5 and 6–10 years. All data were analyzed using the

SAS software (SAS 9.4, SAS Institute Inc., Cary, NC). Rejection level of

H0 was set at p < 0.05.

Structural equation modeling was utilized to estimate above-

ground biomass response to rainfall and time since abandonment

based on the hypothesis that rainfall and time since abandonment had

F IGURE 5 Ordination diagram for three cropland groups (A1998,A1999, and A2000) for the aboveground biomass in 10 yearssuccession using NMDS [Colour figure can be viewed atwileyonlinelibrary.com]

HU ET AL. 5

the potential to directly alter aboveground biomass, taxonomic, and

functional diversity, as well as indirectly through changing the taxo-

nomic and functional diversity of aboveground biomass (Ali

et al., 2019). We used the Chi-square (χ2) test to judge the fit of the

model which has a good fit when 0 ≤ χ2/df ≤ 2 and .05 < p ≤ 1

(Grace, 2006). The SEM's were calculated using AMOS

21 (Arbuckle, 2010).

3 | RESULTS

3.1 | Taxonomic diversity

There were significant differences in taxonomic diversity in the first

year after abandonment (Figure 2a–c). Taxonomic diversity of the

three cropland groups increased during the first 5–6 years, with signif-

icant differences in some years. Species richness (Figure 2a) differed

significantly in the first, sixth, and seventh year. The Shannon–Wiener

index differed significantly in the first, fifth, sixth, and eighth year.

Species evenness differed significantly in the first, third, sixth, and

eighth year. There were no differences in other years. During years

1–5, species richness was only regulated by time since abandonment

(p < 0.05); Shannon–Wiener index was only affected by time of aban-

donment; and species evenness was not affected by any factors

(Figure 2a–c). Rainfall, time since abandonment, time of abandonment,

and their interactions had no effect on species richness and Shannon–

Wiener index at 6–10 years. Species evenness was affected by the

interaction of time since abandonment and time of abandonment

(p < 0.05; Table 1).

3.2 | Functional diversity

Our findings for functional diversity are summarized on the right side

of Figure 3. The MFAD (Figure 2d) of the three cropland groups had

no difference at first 4 years (p > 0.05), but differed significantly in

the fifth, sixth, and eighth years. The MFAD was significantly regu-

lated by time of abandonment (p < 0.01) in the first 5 years but was

not affected by any factors at 6–10 years (Table 1). The functional

divergence (Figure 2e) of all the three cropland groups differed signifi-

cantly in the second, third, and eighth year. It was affected by rainfall

(p < 0.05) and time since abandonment (p < 0.01) in first 5 years, but

it was not affected by any factors in years 6–10 (Table 1). There was a

significant difference in functional evenness (Figure 2f) for the three

cropland groups in the first and fifth year. It was affected by rainfall in

all years of the observation period (10 years), but it was also affected

by time since abandonment from the sixth year (Table 1).

3.3 | Aboveground biomass and stability of thethree cropland groups

The three cropland groups showed a similar trend in aboveground bio-

mass over the study period (Figure 3a). Biomass was greatest at the

beginning of succession and decreased over the next 5–6 years,

exhibiting increased variation with ongoing succession (Figure 3a).

The highest aboveground biomass was reached in the second year for

A1998 and A2000 (297.5 ± 45.0 and 652.8 ± 92.3 g m−2, respec-

tively) and in the first year for A1999 (265.7 ± 12.0 g m−2). Above-

ground biomass varied significantly with rainfall, time of

TABLE 1 The effect of rainfall (R), time since abandonment (Y), time of abandonment (T, A1998, A1999, A2000) and their interactions onaboveground biomass, taxonomic and functional diversity (F and P; *p < 0.05, **p < 0.01, and ***p < 0.001)

Stage VariablesAbovegroundbiomass

Taxonomic diversity Functional diversity

Speciesrichness

Shannon-Wienerindex

Speciesevenness MFAD

Functionaldivergence

Functionalevenness

1–5 years R 9.1** 2.07 0.07 0.01 0.02 2.89* 9.35**

Y 1.38 4.67* 1.24 0.46 0.52 8.11** 0.04

Tr 3.42* 0.16 2.52* 1.79 6.86** 0.95 1.55

R * Y 11.66** 1.44 2.06 0.36 0.98 1.84 0.46

R * T 4.66* 0.14 2.41 1.55 7.83** 1.57 2.71*

Y * T 4.52* 0.15 2.91* 2.12 9.17*** 1.54 4.42*

R * Y * T 6.09** 0.22 2.52* 1.46 9.37*** 2.05 0.46

6–10 years R 5.02* 0.96 0.18 0.25 0.63 0.17 3.20*

Y 3.03 0.93 0.53 1.19 1.26 0.16 4.31*

T 1.12 0.45 1.41 2.61 1.09 1.88 2.32

R * Y 2.80 0.69 0.39 0.83 1.29 0.04 3.70*

R * T 1.24 0.05 0.51 1.38 0.64 1.30 1.32

Y * T 0.82 0.73 2.07 3.55* 1.10 2.05 2.83*

R * Y * T 0.84 0.26 1.17 2.29 0.79 1.53 1.93

aNote: The bold values in the table indicates that the variable has a significant effect at the 0.05, 0.01, or 0.001 level.

6 HU ET AL.

abandonment, and the interactions of three factors in the first 5 years,

but only with rainfall in years 6–10 (Table 1).

The SEM explained 44 and 42% of the variation in aboveground

biomass in years 1–5 and 6–10, respectively (Figure 4). Rainfall con-

tributed a little to aboveground biomass in years 1–5 (Figure 4-a3).

Time since abandonment had a strong direct effect on the above-

ground biomass with standardized path coefficients (spc) of −0.59

(p < 0.01) in years 1–5. Only functional divergence had a strong direct

effect on the aboveground biomass (spc = 0.46, p < 0.05) in 1–5 years

(Figure 4-a1). However, in years 6–10, rainfall had a direct effect on

the aboveground biomass (spc = 0.23, p < 0.05). Time since abandon-

ment had a standardized total effect on aboveground biomass of 0.35,

including direct and indirect effects though functional evenness

(Figure 4-b3).

Stability of aboveground biomass was initially highest in A1998,

but declined during the second period, and then increased in the third

period. Stability of A1999 and A2000 vegetation increased during

over time. There were no differences among the three cropland

groups for the last two periods (Figure 3b).

3.4 | NMDS ordination of the cropland groups

The vegetation composition of three cropland groups was clustered in

the first year following abandonment, but diverged from the second

to fourth year, and subsequently, clustered again (Figure 5). The

annual Salsola collina Pall, a pioneer species, colonized the abandoned

cropland in the first year since abandonment; it accounts for more

than 95% of total biomass in all the three cropland groups (Figure S2

and Table S1).

4 | DISCUSSION

4.1 | Taxonomic and functional diversity

Taxonomic and functional diversity increased rapidly in the first

5 years; subsequently both aspects of diversity stabilized, and these

results are consistent with previous studies (Kou et al., 2015; Liu, Wu,

Ding, Tian, & Shi, 2017). Species richness of the three cropland groups

initially displayed a rapid increase followed but leveled off succession

progressed. This is consistent with other studies, in which the species

richness first increased and then reached a steady state in the vegeta-

tion succession (Bonet & Pausas, 2004; Wang et al., 2009). Prior

observations might explain these dynamic trends. During the early

stages of succession, abandoned croplands have been found to be ini-

tially colonized by fast-growing annual species (Rees, Condit, Crawley,

Pacala, & Tilman, 2001), and Salsola collina Pall is the fast-growing

annual species in this study (Figure S2). Thus, the rate of species gains

or colonization generally overcomes early successional changes

(Anderson, 2007) and may cause a rapid increase in species richness

(Huston, 1997). Subsequently, richness plateaus when the species

gain and loss rates equalize and the communities approach an

equilibrium state (Cardinale et al., 2007; Marquard et al., 2009). In the

later stages of succession, the annual species S. collina failed to com-

pete against long-lived and highly resilient perennial late-successional

dominants and was ultimately replaced by these late species (-

Figure S2), it was consistent with the studies of Pickett (1982) and Liu

et al. (2017).

Both MFAD and Feve increased soon after abandonment, indicat-

ing that the functional traits space more dispersion and species traits

occupied more space in the first 5 years after croplands were aban-

doned. Functional traits and species traits generally became homoge-

neous as succession proceeded, facilitating a more stable community

(Veen et al., 2018). Veen et al. (2018) found that taxonomic and func-

tional diversity changed over time. We found that time of abandon-

ment had a strong effect on the taxonomic diversity and Feve at first

year since abandonment (Figure 2). This indicates that temporal het-

erogeneity can lead to differences in early succession years, especially

taxonomic diversity and Feve. González-Megías, Gómez, and Sánchez-

Pinero (2007) also found that taxonomic diversity was affected by

temporal heterogeneity and controlled by the sampling scale. From

the sixth to the tenth year, temporal heterogeneity had no effect on

taxonomic and functional diversity (Table 1). Temporal heterogeneity

had minimal effect on both taxonomic and functional diversity after

about 8 years (Figure 2) suggesting that time heterogeneity results in

minimal the differences in taxonomic and functional diversity after

this period of time.

4.2 | Aboveground biomass and stability

In this study, biomass in each cropland group was highest soon after

abandonment and declined over time (Figure 3a). Some previous stud-

ies have not reported this pattern (Asefa, Oba, Weladji, &

Colman, 2003; Bonet & Pausas, 2004) while other studies have found

that aboveground biomass may be highest in the early stages of suc-

cession and that this depends on the plant species first occupying the

niche (Clark, Knops, & Tilman, 2018; Zhang, Liu, Xue, & Wang, 2016).

In this study, the first species S. collina has high single-plant biomass,

and that was the reason for high total biomass in the beginning time

since abandonment (Figure S2). The first species in a niche depends

on soil fertility (organic carbon, total nitrogen, nitrate nitrogen, and

available phosphorus, etc.), which in turn depends on prior farming

practices including the use of farmyard manure (Zhang et al., 2016).

Zhang, Ren, et al. (2018b) found strong reciprocity between vegeta-

tion and soil resources. The croplands used in this study were fertil-

ized with farmyard manure which tends to release nutrients slowly,

influencing succession processes including biomass production. Bio-

mass is higher in the early years following abandonment and then

gradually declines in years 5–6, after that the biomass increased

slowly (Figure 3a). That was because the nutrients accumulate in the

soil in part from the decomposition of litter and from root exudates

(Zhang et al., 2016). Furthermore, Jiao, Wen, An, and Yuan (2013) and

Jucker, Bouriaud, Avacaritei, and Coomes (2014) also concluded that

the establishment of plants improved the soil environmental

HU ET AL. 7

conditions. The biomass is relatively stable in all cropland groups after

the fifth year (Figure 3). We believe that the effect of time of aban-

donment on biomass declines over time.

4.3 | Community succession

The annual pioneer species S. collina colonized abandoned cropland

early and became the dominant species in the first few years after

abandonment (Figure S2). Its greater adaptability to the environment

and competitiveness for resources has been shown to negatively

affect the growth of other species (Zhang et al., 2016), consistent with

the low diversity (taxonomic and functional) that we observed at this

stage of succession (Figure 3). But a 3-years period of community

development is too short to enable S. collina to colonize bare soil.

Later more species established (including A. capillaris, S. bungeana, and

L. davurica; Figure S2), and taxonomic and functional diversity were

correspondingly higher (Figure 2).

The vegetation composition was differed among the three

cropland groups from second to fifth years (Figure 5 and Figure S2).

All three cropland groups started with bare land after abandon-

ment, but these differences may be caused by different rainfall

events at the time of abandonment (Table 1). In years 3–8, all three

cropland groups exhibited a change in community structure, with a

large number of new species (Figure S2), and a change in the com-

position of the community (Figure 5). The main reason for this

change may have been the germination of a large number of seeds

in the soil seed bank (Hu, Chen, Chen, & Hou, 2015). However,

because of continuous elimination due to competition, several well

adapted indigenous species (A. capillaris, S. bungeana, L. davurica,

Heteropappus altaicus (Willd.) Novopokr. and Pennisetum cen-

trasiaticum Tzvel.) survived (Hou, Xiao, & Nan, 2002; Zhang, Ren,

et al., 2018b). This is the reason for the similarity of the composi-

tion of the three cropland groups in the 10th year post abandon-

ment (Figure 5 and Figure S2).

Jírová, Klaudisová, and Prach (2012) found that abandoned fields

were situated in a more humid region in the Czech Republic, would

develop into shrublands and ultimately to the climax vegetation

including forest if succession continues for the long term, over many

decades. Our study area was in semiarid region, without natural

woody plants present in the local grasslands, limiting the potential for

woody vegetation (Sojneková & Chytry, 2015). Consequently, it is

highly unlikely that long-term succession would lead to the presence

of woody vegetation, either indigenous or exotic on this region.

Sojneková and Chytry (2015) found that species composition of

restored sites became similar to native dry grassland species after

40 years. On the Loess Plateau of China, Liu et al. (2017) predicted

the abandoned cropland will take at least 20 years to recover to a sta-

ble community structure. Hou et al. (2002) predicted it would take

8–9 years to establish stable community structure and 9–11 years for

the dominant population to be stable. In this research, the three crop-

land groups reached a stable dominant population to native grassland

species by 10 years, consistent with the prediction of Hou et al. (2002)

but faster than observed by Liu et al. (2017). The rainfall and local

vegetation types may be the main factors governing the speed of suc-

cession, we found that the biomass and some diversity indexes were

strongly influenced by rainfall in this study (Table 1).

5 | CONCLUSIONS

During the succession of all three cropland groups, the plant commu-

nity transitioned from dominance by therophytes with relatively high

aboveground biomass to dominance by perennial species with lower

aboveground biomass. Time heterogeneity affected aboveground bio-

mass, taxonomic, and functional diversity in the progress of succes-

sion, especially for biomass and taxonomic diversity in the first year,

but had no effect on biomass and diversity of taxonomic and func-

tional after 8 years. Therefore, we suggest that it is feasible to study

the succession of abandoned cropland beyond 8 years using temporal

heterogeneity. However, because time heterogeneity in our study

was based on croplands abandoned in three consecutive years, we

cannot yet make conclusions about differences in succession between

croplands with a longer interval between abandonment. Ongoing

monitoring will be necessary for future studies of succession in aban-

doned croplands.

ACKNOWLEDGMENTS

We specially thank Prof. James Millner and Saman Bowatte for editing

the English and providing helpful and constructive comments for

improving the manuscript. And we also thank Prof. Xiongzhao He for

the help in statistical analysis. This research was supported by the

Strategic Priority Research Program of the Chinese Academy of Sci-

ences (XDA2010010203), the National Natural Science Foundation of

China (31172249), the Program for Changjiang Scholars and Innova-

tive Research Team in University (IRT17R50), Key R & D Program of

Ningxia Hui Autonomous Region (2019BBF02001).

CONFLICT OF INTEREST

The authors declare no conflicts of interest.

AUTHOR CONTRIBUTIONS

Conceptualization and supervision, Fu-Jiang Hou; Design and meth-

odology, Fu-Jiang Hou and Zhi-Biao Nan; field investigation, Fu-Jiang

Hou, An Hu, Sheng-Hua Chang, and Xian-Jiang Chen; data analysis,

Fu-Jiang Hou and An Hu; data curation, Fu-Jiang Hou and An Hu;

writing—original draft preparation, Fu-Jiang Hou and An Hu; writing—

review and editing, Fu-Jiang Hou. All the authors have read and

agreed to the published version of the manuscript.

DATA AVAILABILITY STATEMENT

The data that support the findings of this study are available from the

corresponding author upon reasonable request.

ORCID

Fu-Jiang Hou https://orcid.org/0000-0002-5368-7147

8 HU ET AL.

REFERENCES

Ali, A., Lin, S., He, J., Kong, F., Yu, J., & Jiang, H. (2019). Elucidating space,

climate, edaphic and biodiversity effects on aboveground biomass in

tropical forests. Land Degradation & Development, 30, 918–927.https://doi.org/10.1002/ldr.3278

Anderson, K. J. (2007). Temporal patterns in rates of community change

during succession. The American Naturalist, 169, 780–793. https://doi.org/10.1086/516653

Arbuckle, J. (2010). IBM SPSS Amos 19 User's guide. Crawfordville, FL:

Amos Development Corporation.

Asefa, D. T., Oba, G., Weladji, R. B., & Colman, J. E. (2003). An assessment

of restoration of biodiversity in degraded high mountain grazing lands

in northern Ethiopia. Land Degradation & Development, 14, 25–38.https://doi.org/10.1002/ldr.505

Benjamin, K., Domon, G., & Bouchard, A. (2005). Vegetation composition

and succession of abandoned farmland: Effects of ecological, historical

and spatial factors. Landscape Ecology, 20, 627–647. https://doi.org/10.1007/s10980-005-0068-2

Bonet, A., & Pausas, J. G. (2004). Species richness and cover along a

60-year chronosequence in old-fields of southeastern Spain. Plant

Ecology, 174, 257–270. https://doi.org/10.1023/B:VEGE.

0000049106.96330.9c

Botta-Dukát, Z. (2010). Rao's quadratic entropy as a measure of functional

diversity based on multiple traits. Journal of Vegetation Science, 16,

533–540. https://doi.org/10.1111/j.1654-1103.2005.tb02393.xCardinale, B. J., Wright, J. P., Cadotte, M. W., Carroll, I. T., Hector, A.,

Srivastava, D. S., … Weis, J. J. (2007). Impacts of plant diversity on bio-

mass production increase through time because of species comple-

mentarity. Proceedings of the National Academy of Sciences of the

United States of America, 104, 18123–18128. https://doi.org/10.

1073/pnas.0709069104

Chen, X. J., Hou, F. J., Matthew, C., & He, X. Z. (2017). Soil C, N, and P

stocks evaluation under major land uses on China's Loess Plateau.

Rangeland Ecology & Management, 70, 341–347. https://doi.org/10.1016/j.rama.2016.10.005

Clark, A. T., Knops, J. M. H., & Tilman, D. (2018). Contingent factors

explain average divergence in functional composition over 88 years of

old field succession. Journal of Ecology, 107, 545–558. https://doi.org/10.1111/1365-2745.13070

Collins, S. L., Avolio, M. L., Gries, C., Hallett, L. M., Koerner, S. E., La

Pierre, K. J., … Jones, M. B. (2018). Temporal heterogeneity increases

with spatial heterogeneity in ecological communities. Ecology, 99,

858–865. https://doi.org/10.1002/ecy.2154Deng, L., Wang, K. B., Zhu, G. Y., Liu, Y. L., Chen, L., & Shangguan, Z. P.

(2018). Changes of soil carbon in five land use stages following

10 years of vegetation succession on the Loess Plateau, China. Catena,

171, 185–192. https://doi.org/10.1016/j.catena.2018.07.014González-Megías, A., Gómez, J. M., & Sánchez-Pinero, F. (2007). Diversity-

habitat heterogeneity relationship at different spatial and temporal

scales. Ecography, 30, 31–41. http://doi.org/10.2307/30243195Grace, J. B. (2006). Structural equation modeling and natural systems. Cam-

bridge, UK: Cambridge University Press.

Hodapp, D., Borer, E. T., Harpole, W. S., Lind, E. M., Seabloom, E. W.,

Adler, P. B., … Hillebrand, H. (2018). Spatial heterogeneity in species

composition constrains plant community responses to herbivory and

fertilization. Ecology Letters, 21, 1364–1371. https://doi.org/10.1111/ele.13102

Hou, F. J., Xiao, J. Y., & Nan, Z. B. (2002). Eco-restoration of abandoned

farmland in the Loess Plateau. Chinese Journal of Applied Ecology, 13,

923–929. https://doi.org/10.13287/j.1001-9332.2002.0216.Hu, A., Chen, H., Chen, X. J., & Hou, F. J. (2015). Soil seed banks of crop-

land and rangeland on the Loess Plateau. Pratacultural Science, 32,

1035–1040. https://doi.org/CNKI:SUN:CYKX.0.2015-07-001.

Hu, A., Zhang, J., Chen, X. J., Chang, S. H., & Hou, F. J. (2019). Winter graz-

ing and rainfall synergistically affect soil seed bank in a semiarid area.

Rangeland Ecology and Management, 72, 160–167. https://doi.org/10.1016/j.rama.2018.07.012

Huston, M. A. (1997). Hidden treatments in ecological experiments: Re-

evaluating the ecosystem function of biodiversity. Oecologia , 110,

449–460. https://doi.org/10.1007/s004420050180Jiao, F., Wen, Z. M., An, S. S., & Yuan, Z. (2013). Successional changes in

soil stoichiometry after land abandonment in Loess Plateau, China.

Ecological Engineering, 58, 249–254. https://doi.org/10.1016/j.

ecoleng.2013.06.036

Jírová, A., Klaudisová, A., & Prach, K. (2012). Spontaneous restoration of

target vegetation in old-fields in a Central European landscape: A

repeated analysis after three decades. Applied Vegetation Science, 15,

245–252. https://doi.org/10.1111/j.1654-109X.2011.01165.xJucker, T., Bouriaud, O., Avacaritei, D., & Coomes, D. A. (2014). Stabilizing

effects of diversity on aboveground wood production in forest ecosys-

tems: Linking patterns and processes. Ecology Letters, 17, 1560–1569.https://doi.org/10.1111/ele.12382

Kou, M., Jiao, J., Yin, Q., Wang, N., Wang, Z. J., Li, Y. J., … Cao, B. T. (2015).

Successional trajectory over 10 years of vegetation restoration of

abandoned slope croplands in the hill-gully region of the Loess Pla-

teau. Land Degradation & Development, 27, 1–14. https://doi.org/10.1002/ldr.2356

Limb, R. F., Hovick, T. J., Norland, J. E., & Volk, J. M. (2018). Grassland

plant community spatial patterns driven by herbivory intensity. Agricul-

ture, Ecosystems & Environment, 257, 113–119. https://doi.org/10.

1016/j.agee.2018.01.030

Liu, Y., Wu, G. L., Ding, L. M., Tian, F. P., & Shi, Z. H. (2017). Diversity-

productivity trade-off during converting cropland to perennial grass-

land in the semi-arid areas of China. Land Degradation & Development,

28, 699–707. https://doi.org/10.1002/ldr.2561Marquard, E., Weigelt, A., Temperton, V. M., Roscher, C., Schumacher, J.,

Buchmann, N., … Schmid, B. (2009). Plant species richness and func-

tional composition drive overyielding in a six-year grassland experi-

ment. Ecology, 90, 3290–3302. https://doi.org/10.1890/09-0069.1Mcgranahan, D. A., Engle, D. M., Fuhlendorf, S. D., Winter, S. J.,

Miller, J. R., & Debinski, D. M. (2012). Spatial heterogeneity across five

rangelands managed with pyric-herbivory. Journal of Applied Ecology,

49, 903–910. https://doi.org/10.1111/j.1365-2664.2012.02168.xMcIntosh, R. P. (1991). Concept and terminology of homogeneity and het-

erogeneity in ecology. In J. Kolasa & S. T. A. Pickett (Eds.), Ecological

heterogeneity (pp. 24–46). New York, NY: Springer-Verlag.

Pickett, S. T. A. (1982). Population patterns through twenty years of old

field succession. Plant Ecology, 49, 45–59. https://doi.org/10.1007/BF00051566

Rees, M., Condit, R., Crawley, M., Pacala, S., & Tilman, D. (2001). Long-

term studies of vegetation dynamics. Science, 293, 650–655. https://doi.org/10.3417/2012016

Ren, J. Z., Hu, Z. Z., Zhao, J., Zhang, D. G., Hou, F. J., Lin, H. L., & Mu, X. D.

(2008). A grassland classification system and its application in China.

The Rangeland Journal, 30, 199–209. https://doi.org/10.1071/

RJ08002

Schmera, D., Er}os, T., & Podani, J. (2009). A measure for assessing func-

tional diversity in ecological communities. Aquatic Ecology, 43,

157–167. https://doi.org/10.1007/s10452-007-9152-9Smith, C. C. (1940). Biotic and physiographic succession on abandoned

eroded farmland. Ecological Monographs, 10, 421–484. https://doi.org/10.2307/1948513

Sojneková, M., & Chytry, M. (2015). From arable land to species-rich semi-

natural grasslands: Succession in abandoned fields in a dry region of

Central Europe. Ecological Engineering, 77, 373–381. https://doi.org/10.1016/j.ecoleng.2015.01.042

Thomas, C. A., Knops, J. M. H., & Dave, T. (2018). Contingent factors

explain average divergence in functional composition over 88 years of

old field succession. Journal of Ecology, 107, 545–558. https://doi.org/10.1111/1365-2745.13070

HU ET AL. 9

Veen, G. F., Van Der Putten, W. H., & Bezemer, T. M. (2018). Biodiversity-

ecosystem functioning relationships in a long-term non-weeded field

experiment. Ecology, 99, 1836–1846. https://doi.org/10.1002/ecy.

2400

Villéger, S., Mason, N. W. H., & Mouillot, D. (2008). New multidimensional

functional diversity indices for a multifaced framework in functional

ecology. Ecology, 89, 2290–2301. https://doi.org/10.1890/07-1206.1Wang, G., Liu, G., & Xu, M. (2009). Above- and belowground dynamics of

plant community succession following abandonment of cropland on

the Loess Plateau, China. Plant and Soil, 322, 343–343. https://doi.org/10.1007/s11104-008-9773-3

Xu, L. (2015). Environmental adaptability and productivity of agroecological

system in the Longdong Loess Plateau (pp. 43–47). Lanzhou: LanzhouUniversity.

Xu, B. C., Xu, W. Z., & Huang, J. (2011). Biomass allocation, relative com-

petitive ability and water use efficiency of two dominant species in

semiarid Loess Plateau under water stress. Plant Science, 181,

644–651. https://doi.org/10.1016/j.plantsci.2011.03.005Zhang, C., Liu, G., Song, Z., Qu, D., Fang, L., & Deng, L. (2017a). Natural

succession on abandoned cropland effectively decreases the soil erod-

ibility and improves the fungal diversity. Ecological Applications, 27,

2142–2154. https://doi.org/10.1002/eap.1598Zhang, C., Liu, G. B., Xue, S., & Wang, G. L. (2016). Soil bacterial commu-

nity dynamics reflect changes in plant community and soil properties

during the secondary succession of abandoned farmland in the Loess

Plateau. Soil Biology & Biochemistry, 97, 40–49. https://doi.org/10.

1016/j.soilbio.2016.02.013

Zhang, K. R., Cheng, X. L., Xiao, S., Liu, Y., & Zhang, Q. F. (2018a). Linking

soil bacterial and fungal communities to vegetation succession follow-

ing agricultural abandonment. Plant and Soil, 431, 19–36. https://doi.org/10.1007/s11104-018-3743-1

Zhang, W., Ren, C. J., Deng, J., Zhao, F. Z., Yang, G. H., Han, X. H., …Feng, Y. Z. (2018b). Plant functional composition and species diversity

affect soil C, N, and P during secondary succession of abandoned

farmland on the Loess Plateau. Ecological Engineering, 122, 91–99.https://doi.org/10.1016/j.ecoleng.2018.07.031

Zhang, Y., Yu, M., Guo, J., Wu, D., Hua, Z. S., & Lu, H. (2017b). Spatiotem-

poral heterogeneity of core functional bacteria and their synergetic

and competitive interactions in denitrifying sulfur conversion-assisted

enhanced biological phosphorus removal. Scientific Reports, 7, 10927.

https://doi.org/10.1038/s41598-017-11448-x.

Zhong, Y. Q. W., Yan, W. M., Wang, R. W., Wang, W., & Shangguan, Z. P.

(2018). Decreased occurrence of carbon cycle functions in microbial

communities along with long-term secondary succession. Soil Biology &

Biochemistry, 123, 207–217. https://doi.org/10.1016/j.soilbio.2018.

05.017

Valkó, O., Deák, B., Török, P., Kelemen, A., Miglécz, T., Tóth, K.,

Tóthmérész, B. (2016). Abandonment of croplands: problem or chance

for grassland restoration? case studies from hungary. Ecosystem Health

and Sustainability, 2, (2), e01208. http://dx.doi.

org/10.1002/ehs2.1208.

SUPPORTING INFORMATION

Additional supporting information may be found online in the

Supporting Information section at the end of this article.

How to cite this article: Hu A, Chang S-H, Chen X-J, Hou F-J,

Nan Z-B. Temporal heterogeneity has no effect on the

direction of succession in abandoned croplands in a semiarid

area of northwest China. Land Degrad Dev. 2020;1–10.

https://doi.org/10.1002/ldr.3780

10 HU ET AL.