Advertisement

Journal of Plant Research

, Volume 132, Issue 3, pp 345–358 | Cite as

Interaction of livestock grazing and rainfall manipulation enhances herbaceous species diversity and aboveground biomass in a humid savanna

  • Daniel Osieko OkachEmail author
  • Joseph O. Ondier
  • Gerhard Rambold
  • John Tenhunen
  • Bernd Huwe
  • Eun Young Jung
  • Dennis O. Otieno
Regular Paper

Abstract

Understanding of the interaction of livestock grazing and rainfall variability may aid in predicting the patterns of herbaceous species diversity and biomass production. We manipulated the amount of ambient rainfall received in grazed and ungrazed savanna in Lambwe Valley-Kenya. The combined influence of livestock grazing and rainfall on soil moisture, herbaceous species diversity, and aboveground biomass patterns was assessed. We used the number of species (S), Margalef’s richness index (Dmg), Shannon index of diversity (H), and Pileou’s index of evenness (J) to analyze the herbaceous community structure. S, Dmg, H and J were higher under grazing whereas volumetric soil water contents (VWC) and aboveground biomass (AGB) decreased with grazing. Decreasing (50%) or increasing (150%) the ambient rainfall by 50% lowered species richness and diversity. Seasonality in rainfall influenced the variation in VWC, S, Dmg, H, and AGB but not J (p = 0.43). Overall, Dmg declined with increasing VWC. However, the AGB and Dmg mediated the response of H and J to the changes in VWC. The highest H occurred at AGB range of 400–800 g m−2. We attribute the lower diversity in the ungrazed plots to the dominance (relative abundance > 70%) of Hyparrhenia fillipendulla (Hochst) Stapf. and Brachiaria decumbens Stapf. Grazing exclusion, which controls AGB, hindered the coexistence among species due to the competitive advantage in resource utilization by the more dominant species. Our findings highlight the implication of livestock grazing and rainfall variability in maintaining higher diversity and aboveground biomass production in the herbaceous layer community for sustainable ecosystem management.

Keywords

Herbaceous layer community Plant biomass Rainfall variability Savanna ecosystem Species evenness 

Introduction

Grazing, fire, soil moisture, and nutrient availability are the major modifiers of the vegetation structure and function in the African savannas (Cumming 1982; Hulme et al. 2001). Extreme weather events due to climate change, expanding agriculture and settlement resulting from increased human population threaten the sustainability of these ecosystems. The areas affected most include the East African and Sudanian savannas (Case 2006; Davidson et al. 2003; Osborne et al. 2018), where pastoralism is common. Human and climatic dynamics negatively affect ecosystem services, by decreasing productivity and intensifying loss of biodiversity (Alley et al. 2003; Anderson et al. 2007). The extent to which such changes affect the herbaceous diversity and productivity in the East Africa savanna is still unclear.

Grazing modifies savanna ecosystems through the action of herbivores on vegetation and soil (Augustine and McNaughton 2006; Harrison and Bardgett 2004; Savadogo et al. 2008). Livestock grazing, for instance, adds nutrients to the soil through their dung and urine (Bardgett and Wardle 2003; Olff and Ritchie 1998), clip the vegetation, and compact the soils by trampling. Through selective removal of the herbaceous vegetation, livestock grazing influence biomass and species richness (Kioko et al. 2012). The response of the herbaceous species diversity to grazing, however, strongly depends on the grazing intensity (Adler et al. 2001; Graham and Duda 2011). This results from the trade-off between the ability of plants to withstand grazing pressure, enhance their utilization of resources, and increase their growth patterns (Bakker et al. 2003). For instance, at lower grazing intensity, herbaceous diversity increases due to competitive exclusion which allows the less dominant plants to access vital resources (Borer et al. 2014; Hanke et al. 2014; Koerner and Collins 2014). However, conflicting reports exist on the effects of intense grazing on the herbaceous communities (Cingolani et al. 2005). Overgrazing, therefore, decreases or has no effect on the herbaceous diversity (Anderson et al. 2007; Cingolani et al. 2005; Frank 2005; Milchunas et al. 1988). Such patterns result from the increased functional redundancy of the grazing-tolerant species (Milchunas et al. 1988; Mouillot et al. 2013), modification of nutrient hotspots (Marshall et al. 2018; van der Waal et al. 2011), and the adaptability of ecosystems to long term grazing pressure (Cingolani et al. 2005; Marshall et al. 2018; Savadogo et al. 2008).

In East Africa, models predict a 5–10% decrease and 5–20% increase in rainfall in dry and wet seasons, respectively by 2050 (Conway 2009; Hulme et al. 2001), which is likely to affect ecosystem structure and functions. Rainfall manipulation experiments at plot scale have been used to predict how changes in precipitation affect the natural ecosystems (Jones et al. 2016; Knapp et al. 2008; Nippert et al. 2006; Swemmer et al. 2007). Such studies assume that the influences of precipitation on the vegetation are primarily a function of soil moisture. However, temperature and light intensity also vary with rainfall events and are therefore significant in predicting precipitation patterns in the natural ecosystems (Zeppel et al. 2014). Despite the projected rainfall variability in East Africa (Case 2006), the application of rainfall manipulation experiments to monitor the effects of extreme precipitation is largely missing (Beier et al. 2012). There is no quantified effect of extreme rainfall events on the spatial and temporal heterogeneity of the herbaceous vegetation within these ecosystems. Such findings will provide insight on the potential impact of increased rainfall variability on the herbaceous diversity and biomass production.

In the savannas, seasonality in rainfall results in distinct dry and wet periods, dictating the soil moisture availability, plant physiology, patterns of biomass development, and species richness (William and Albertson 2004; Zeppel et al. 2014). Previous studies have assessed the response of photosynthesis, respiration, and phenology to seasonal variation in soil moisture in the savannas (K’Otuto et al. 2014; Merbold et al. 2009; William and Albertson 2004). These studies underscore the importance of soil moisture availability on the functionality of grazed and non-grazed savanna ecosystems. However, the effect of soil moisture variation on the herbaceous species diversity of these ecosystems remains unclear, therefore, necessitating further investigations. Studies conducted under ambient rainfall across East Africa (Augustine 2003; Harrison and Bardgett 2004; Kioko et al. 2012) and in the Lambwe Valley ecosystem in Kenya (K’Otuto et al. 2012; Otieno et al. 2011) have reported variations in herbaceous biomass in response to seasonal changes in soil moisture. These studies concluded that changes in the herbaceous biomass were more affected by the seasonal and not annual rainfall totals.

Grazing intensity is rising in Lambwe Valley, but its effect on the ecosystem has not been fully explored. K’Otuto et al. (2012) and Otieno et al. (2010, 2011) examined the effects of cattle grazing on CO2 exchange and biomass production in the herbaceous layer community of this ecosystem. These studies reported significant reduction in herbaceous productivity due to grazing. To the best of our knowledge, the interactive influence of livestock grazing and rainfall intensity on the structure and function of herbaceous community in Lambwe Valley, remains unexplored. However, in other ecosystems, livestock grazing modified the effect of rainfall on soil moisture availability, nutrient distribution, and herbaceous species diversity (Dangal et al. 2016; Koerner and Collins 2014; Porensky et al. 2013). The interaction of grazing and other climatic factors additionally facilitates such changes. We selected Lambwe Valley as a typical savanna to test the interactive influence of livestock grazing and rainfall variability on species diversity and aboveground biomass production. This ecosystem was suitable because of its humid nature characterized by higher rainfall variability and the large population of livestock that dominate the area. The current study assessed how grazing and manipulated rainfall affected the herbaceous species diversity and aboveground biomass production (AGB) in a typical savanna ecosystem in Lambwe Valley Kenya. We hypothesized that: (1) herbaceous diversity shifts seasonally with rainfall amount. The changes are, however modified by livestock grazing; (2) higher AGB increases the herbaceous community diversity through enhanced soil moisture condition.

Materials and methods

Study area

The study was conducted in the Lambwe Valley (00°35′27.72′′S and 34°18′81.64′′E), a humid savanna covering 324 km2 and hosts Ruma National Park (120 km2) at an altitude of 1200–1600 m a.s.l. It is located 10 km east of Lake Victoria, southeast of Homa-Bay town and East of Gembe and Gwasi Hills in Homa-Bay County, Kenya. The annual rainfall is 900–1,300 mm in two seasons: April–June (long rains) and October–December (short rains) with mean air temperature of 25 ºC. In the last 15 years, however, rainfall has significantly decreased at a higher interannual variability characterized by a coefficient of variation (CV) of 20.07%. The rainy seasons that initially began in mid-March and mid-September are currently, delayed until April and October, respectively. During the study period, the rainfall amount realized (945.6 mm) was below the mean of the last 15 years (1,085 mm). The soils at the valley bottom are Vertisols with biotite rock (Allsopp and Baldry 1972). Hyparrhenia fillipendula, with thickets of Acacia and other shrubs dominate the landscape. Peasant farming and livestock keeping are the common economic activities in the area. Measurements were done in selected open areas neighboring Ruma National Park, on land belonging to the National Youth Service (NYS).

Experimental design

The design was split plot factorial with 2 levels of grazing: grazed (G) and ungrazed (U) and 3 levels of rainfall manipulation: ambient rainfall (100%), 50% reduction (50%), and 50% increase (150%) of the ambient rainfall. The setup was replicated three times, giving a total of 18 plots. The plots were abbreviated using the first letters of the grazing levels (G or U) followed by rainfall amount (50, 100, and 150%) i.e. G50%, U150%, etc. Initially grazed areas were fenced in 2013 (2 m high) to exclude livestock (goats and cattle) (ungrazed). Rainfall manipulation plots (6 m × 3 m wide and 2 m high) were constructed in 2015 (3 months before the first measurement) to exclude and include the desired amount of ambient rainfall. To achieve 50% rainfall reduction, rain exclusion gutters from strips of transparent plastic (3 mm XT type 20070; light transmission 95%; 0.2 °C temperature reduction) sheets (30 cm width) were intermittently stretched to cover 50% of the total plot/ground surface (10 strips were used). The strips were laid (14° inclined) and stretched over 2 m high metal frames to allow free movement of livestock over the grazed plots. The excluded rainwater was redirected on to the 150% rain treatment plots, giving 150% of the ambient rainfall (Fig. 1). Trenches (50 cm deep and 30 cm wide), reinforced with plastic sheets buried in the soil, were dug between plots to reduce lateral water flow across plots. Control plots were uncovered to receive ambient rainfall. The shelters had open sides to reduce the greenhouse effect from gutters. Routinely, livestock were released into the grazing field at 11:00 am and driven back to their pen at 5:00 pm. Their pattern of grazing was random with our plots having equal chances of being grazed as their surroundings. Trampling, defoliation, dunging and urination were common features associated with grazing.
Fig. 1

Layout of the rainfall manipulation plots (a) separated by trenches (50 cm deep and 30 cm wide) shielded by polythene to prevent lateral flow of water. The grey parallel arrows show the flow of 50% rainwater to enhance the plot designed to receive 150% of the total amount of rainfall through gravity. Rainfall manipulation plot (b) partially covered with intermittently spread polythene sheets on a slanting roof (14°) to aid the flow of rainwater. Livestock grazing (c) within the rainfall manipulation plots and their surroundings

Microclimate and soil volumetric water content (VWC)

Rainfall, air temperature, humidity, and light intensity were logged from three automatic microclimate stations (AWS-WS-GP1, Delta-T Devices, Cambridge, UK), installed 2 m aboveground, at the study site. VWC was measured using 5TE probes connected to EM50 data loggers (Decagon Devices Inc., Washington, USA) at 20 cm soil depth. Three probes were installed in each plot to log data at 30 min’ interval between 15/9/2015 and 2/6/2016.

Herbaceous community assessment

We monitored the herbaceous species seasonally, in September 2015, January 2016, and May 2016. The species ground cover within plots (1 m × 1 m) was determined using point frame method (Bonham 1989). The pinholes on the frame were set 10 cm apart. The relative species abundance was calculated from the total count of individuals as outlined in Table 1. To assess changes in richness, diversity, and evenness, we used the Margalef’s richness index (Dmg), Shannon’s diversity (H), and Pielou’s evenness (J), respectively (Magurran 2004). The indices were computed from abundance data using the respective formula in Table 1.
Table 1

Formula for calculating species diversity parameters

Index

Formula

Ground cover

\(Cover = \frac{number\;of\;hits\;of\;species\;x}{total\;numberof\;hits\;forallspecies} \times 100\)

Margalef index of richness

\(D_{Mg} = \frac{{\left( {S - 1} \right)}}{1nN}\)

Shannon index of diversity

\(H = - \mathop \sum \nolimits \left( {\frac{ni}{N}} \right) 1n \left( {\frac{ni}{N}} \right) {\text{or }} - \mathop \sum \nolimits Pi 1n Pi\)

Pielou’s evenness

J = H/1nS

S total number of plant species, N total number of individuals, 1n natural log, ni important value index of the ith species

Aboveground biomass (AGB)

AGB was sampled both destructively and non-destructively from every plot. Two plots (40 × 40 cm) were demarcated on opposite sides of every 1 × 1 m, vegetation monitoring plots for destructive and non-destructive estimation of AGB, respectively. A total of 18 plots were used for destructive sampling of AGB. We used the photographic method for non-destructive estimation of biomass using a digital camera (Tackenberg 2007). The photographic images were calibrated using AGB destructively sampled from selected plots at different times. The validation was done by comparing the exact biomass destructively sampled and their corresponding pixels from the images taken, using linear regression model. Subsequently, the vertical distribution of plant matter in terms of characterized pixels, was evaluated. The images were shot by high-resolution digital camera (20.4 mega-pixels) at 0.5–1.5 m height (depending on the height of the herbaceous stand) against a dark background with a scale. The standing plant mass was then harvested at the ground level and weighed to quantify the fresh AGB. The two most dominant species from the grazed and ungrazed plots were sorted from the harvested samples and their masses established relative to the total AGB. The samples were later oven dried at 80 °C for 48 h. The AGB estimations in the subsequent seasons were done using the non-destructive method. All images were saved in JPG format for analysis using Image J 1.46 software (Schneider et al. 2012).

Statistical analysis

The effects of grazing, rainfall manipulation, seasonality in rainfall, and their interaction on VWC, AGB, S, Dmg, H and J were tested using factorial ANOVA (fully crossed) at a significance level of p ≤ 0.05. The possible interaction of factors tested included: Grazing X rainfall manipulation; grazing X seasonality in rainfall; rainfall manipulation X seasonality in rainfall and grazing X rainfall manipulation X seasonality in rainfall. Owing to the large dataset, the real-time VWC (logged every 30 min throughout the measurement period) was analyzed separately using repeated measure ANOVA. We aggregated the VWC data sets into daily (to minimize the effect of peak VWC values resulting from higher rainfall intensity) and seasonal (September, January and May) means (± SD) to facilitate the subsequent analysis. Generalized linear mixed model (GLMM) assuming a fixed effect was used to test the influence of VWC as a fixed variable on the AGB, S, Dmg, H, and J. The multiple comparison of means was performed using Tukey HSD, when the ANOVA was significant. These analyses were done using JMP 14 Statistical software, SAS Inc.

Complex relationships amongst VWC, biomass, and species diversity parameters were evaluated using structural equation modelling (SEM) (Grace et al. 2010). The direct and indirect (with mediation) effects of VWC on herbaceous biomass and diversity were determined using path analysis in the AMOS 25 software (Arbuckle 2017). Out of the many models tested to predict the effect of VWC on herbaceous biomass and diversity, only one, having Dmg, and AGB as mediators passed the model fit test. We tested the models’ fit using the Chi square goodness of fit (Chi square =2.0–5.0; p > 0.05), confirmatory fit index (CFI > 0.95), and root mean square error of approximation (RSMEA< 0.05). Subsequently, covariances that displayed stronger relationships were introduced to modify the model. The Preacher and Hayes (2008) approach using a bootstrap test with bias corrected (BC) percentile method (95%) was applied to demonstrate mediation. Subsequently regression analysis was conducted using JMP 14 Statistical software to test the extent to which AGB influenced H.

Results

Effect of rainfall manipulation and grazing on soil water status

The pattern of soil VWC at the 20 cm depth corresponded to the changes in rainfall amount during the measurement period (Figs. 2, 3a, b). The VWC increased with the amount of rainfall allowed into the plots along a gradient of 50, 100, and 150% (Fig. 3a, b; Table 2). The lowest mean VWC for the period was 9.78% ± 0.82 in September 2015 whereas, the highest was 27.75% ± 0.88 in May 2016 from G50% and U150% plots, respectively. Grazing had a lower mean effect on VWC (17.46% ± 3.86), compared to ungrazed (22.31% ± 5.00) (Fig. 3b). Rainfall manipulation significantly (f = 96.72; p< 0.05) influenced VWC in both the grazed and ungrazed plots (Fig. 3b). Naturally, VWC varied seasonally and was lowest in September and highest in May, corresponding to the dry and wet seasons, respectively (Fig. 3b). During the dry season, VWC varied significantly between 50, 100 and 150% plots of the ungrazed sites, but not in the grazed site (Fig. 3b). In the rainy season, there were no differences in VWC between 100 and 150% plots of grazed and ungrazed sites. However, the 50% plots had lower VWC. Both grazing and seasonality in rainfall significantly (p< 0.05) interacted with rainfall manipulation to alter the VWC (Table 2).
Fig. 2

Patterns of air temperature and rainfall from the microclimate station at the study site beginning January 2015–September 2016. The unshaded bars highlighted between the two arrows pointing downwards represent the study period

Fig. 3

Pattern of the mean daily volumetric water content of grazed and ungrazed plots for the entire measurement period (a) and its seasonal variations (b) among the treatment plots (50%, 100%, and 150%). Different letters show significant variation in means (p < 0.05) among treatments. The error bars in the graph represent mean ± SD

Table 2

Full factorial repeated measures ANOVA, F statistics with P values (in parenthesis, bold values indicate significance) showing how various parameters are influenced by grazing—G (grazed and ungrazed), rainfall manipulation—RMP (50, 100, and 150%) across 3 seasons (dry season of September 2015; wet seasons of January and May 2016) and the interaction of the three variables

Parameters

df

VWC (Volumetric water content)

S (Number of species)

Dmg (Margalef’s index of richness)

H (Shannon’s index of diversity)

J (Pielou’s evenness)

AGB (Aboveground biomass)

Grazing (G)

1

154.12 (< 0.05)

199.55 (< 0.05)

267.85 (< 0.05)

277.98 (< 0.05)

190.96 (< 0.05)

385.40 (< 0.05)

RMP

2

96.72 (< 0.05)

256.64 (< 0.05)

238.97 (< 0.05)

63.57 (< 0.05)

12.83 (< 0.05)

157.76 (< 0.05)

Grazing*RMP

2

14.46 (< 0.05)

6.50 (0. 05)

7.37 (0. 05)

13.12 (< 0.05)

10.57 (0. 05)

18.88 (< 0.05)

Seasonality in rain

2

114.58 (< 0.05)

16.99 (< 0.05)

19.70 (< 0.05)

4.74 (0.01)

0.88 (0.43)

110.78 (< 0.05)

Grazing*Season

2

1.47 (0.24)

4.756 (0.01)

1.97 (0.15)

0.23 (0.79)

0.16 (0.85)

15.61 (< 0.05)

RMP*Season

4

2.74 (0.04)

1.59 (0.20)

1.25 (0.31)

1.05 (0.40)

0.54 (0.71)

7.13 (0. 05)

G*RMP*Season

4

0.95 (0.45)

9.45 (< 0.05)

9.27 (< 0.05)

0.85 (0.50)

0.14 (0.97)

10.04 (< 0.05)

VWC

 

1.65 (0.21)

4.26 (0.04)

8.18 (0.01)

12.75 (0.05)

147.81 (< 0.05)

The last row outlines the outcome of one-way ANOVA indicating the subsequent effect of VWC on the measured parameters

Herbaceous vegetation response to rainfall manipulation and grazing

Table 3 shows the herbaceous species composition and their distribution in the treatment plots. Hyparrhenia fillipendulla (Hochst) Stapf. and Brachiaria decumbens Stapf. dominated the ungrazed plots whereas Bothriochloa insculpta (A. Rich) A. Camus and Paspalum dilatatum Poir were dominant in the grazed plots. The highest number of species (S) occurred in the G100% and U100% plots (Table 3).
Table 3

Relative abundance (mean and SD) showing the distribution of plant species present in plots from grazed and ungrazed areas sampled in September (2015), January and May (2016)

 

Plant species

Ground cover (%)

G50%

G100%

G150%

U50%

U100%

U150%

1

Bothriochloa insculpta (A. Rich) A. Camus

31.11 ± 3.97

27.59 ± 3.68

27.32 ± 4.82

1.11 ± 0.96

2

Paspalum dilatatum Poir

11.87 ± 3.75

18.61 ± 3.48

23.02 ± 4.87

0.62 ± 0.99

0.41 ± 0.62

0.43 ± 0.65

3

Brachiaria decumbens Stapf

2.51 ± 1.61

3.50 ± 1.04

1.86 ± 0.78

27.72 ± 5.54

18.63 ± 4.61

28.75 ± 8.14

4

Hyparrhenia fillipendula (Hochst) Stapf.

12.46 ± 3.46

10.59 ± 2.66

10.82 ± 4.68

54.06 ± 8.30

43.33 ± 9.58

56.47 ± 10.02

5

Acacia spp.

0.39 ± 0.59

6

Agrostis palustris Huds

0.30 ± 0.60

7

Aspilia pluriseta Schweinf

3.26 ± 1.62

2.69 ± 2.75

1.72 ± 0.46

2.43 ± 1.56

4.78 ± 2.00

1.88 ± 1.09

8

Berlaria acanthoides Vahl.

1.49 ± 1.14

1.15 ± 1.01

1.69 ± 1.23

1.61 ± 0.41

0.68 ± 1.07

1.07 ± 1.01

9

Cajanus cajan L. Millsp

0.80 ± 1.30

0.36 ± 1.09

1.09 ± 1.14

0.74 ± 1.41

0.36 ± 0.54

10

Cynodon dactylon (L) Pers

2.35 ± 0.85

1.28 ± 0.54

1.92 ± 0.97

1.78 ± 0.89

0.32 ± 0.97

11

Desmodium gangeticum (L.) D.C.

0.41 ± 0.62

0.93 ± 1.49

1.47 ± 1.05

1.85 ± 1.47

0.37 ± 1.11

12

Digitaria sanguinalis (L) Scop

1.86 ± 0.34

1.88 ± 0.91

1.59 ± 0.41

0.40 ± 0.85

0.74 ± 1.56

13

Euphorbia hirta Linn.

0.71 ± 1.07

1.44 ± 0.25

1.32 ± 1.47

0.96 ± 0.96

0.40 ± 0.85

0.40 ± 0.85

14

Hoslundia opposita Vahl.

1.26 ± 0.53

1.80 ± 1.11

15

Hypoestes aristata Soland ex Roem & Schalt

0.77 ± 1.18

0.68 ± 0.90

0.69 ± 1.10

1.24 ± 1.2

0.75 ± 1.24

16

Hypoestes forskaolii (Vahl) R.Br.

0.45 ± 0.68

1.3 ± 1.11

0.43 ± 0.65

0.81 ± 0.62

17

Indigofera arecta Hochst ex. A. Roch.

1.44 ± 0.25

1.32 ± 1.10

1.20 ± 0.92

18

Indigofera brevicalyx Bak.

0.37 ± 0.97

0.97 ± 0.75

0.43 ± 0.65

19

Ipomoea tenuirostris Steud ex Choisy

4.32 ± 2.03

1.49 ± 0.83

0.41 ± 0.62

3.09 ± 1.23

1.38 ± 1.12

1.18 ± 1.12

20

Justicia striata Vahl

4.38 ± 2.03

3.03 ± 1.60

3.27 ± 1.82

2.61 ± 1.72

1.62 ± 1.65

1.48 ± 1.84

21

Lantana triifolia L.

0.35 ± 1.06

0.96 ± 0.82

22

Leonotis nepetifolia (L) R.Br.

1.21 ± 0.62

23

Ocimmun kilimandscharicum Guerke

0.58 ± 0.87

1.10 ± 0.43

24

Panicum maximum Jacq.

1.10 ± 0.43

25

Rhynchosia minica (L.) DC.

0.86 ± 0.65

0.99 ± 0.75

1.53 ± 0.91

1.07 ± 0.84

26

Sida acuta Burm. F

1.62 ± 0.69

1.29 ± 0.87

0.77 ± 0.96

27

Solanum incanum Linn

1.44 ± 0.25

0.73 ± 1.19

28

Sonchus schweinfurthii Oliv.

0.85 ± 0.89

0.55 ± 0.83

29

Sphaeranthus suaveolens (Forsk) DC

1.44 ± 0.25

0.38 ± 0.80

30

Sporobolus agrostoides Chiov.

10.86 ± 3.10

4.63 ± 2.38

12.85 ± 3.29

0.43 ± 0.65

31

Striga asiatica (L) Kuntze

0.62 ± 0.96

0.45 ± 0.68

0.41 ± 0.62

32

Themeda triandra Forssk

0.58 ± 0.87

0.77 ± 1.36

0.80 ± 0.85

1.09 ± 1.41

33

Triumphetta rhomboidae Jacq.

1.62 ± 0.69

0.98 ± 1.20

0.48 ± 0.73

1.27 ± 0.82

3.18 ± 1.76

2.08 ± 0.95

34

Urena lobata L.

1.47 ± 0.88

1.44 ± 0.25

0.41 ± 0.62

1.19 ± 0.85

0.36 ± 0.54

35

Vernonia glabra (Steetz) Vatke

5.09 ± 2.24

5.42 ± 2.44

4.56 ± 2.12

4.86 ± 2.15

36

Waltheria indica Bak.

1.62 ± 0.69

1.04 ± 1.07

0.75 ± 1.25

The first four species represent the two most numerically dominant plant species, each from grazed (Bothriochloa insculpta and Paspalum dilatatum) and ungrazed (Brachiaria decumbens and Hyparrhenia fillipendula)

Rainfall addition (150%) and reduction (50%) lowered S, Dmg, H, and J in grazed and ungrazed plots (Fig. 4). Higher S and Dmg were reported in the 100% rainfall plot compared to the 50% and 150% plots in both grazed and ungrazed sites (Fig. 4a, b). H and J displayed a similar pattern, only in the ungrazed site, given that, no mean differences were observed among the grazed plots (Fig. 4c, d). The means of S, Dmg, H, and J between the 50 and 150% plots in both grazed and ungrazed sites did not vary. Livestock grazing, rainfall manipulation and seasonality in rainfall influenced the changes in S, Dmg, and H (Table 2). Various interactions especially between grazing and rainfall manipulation additionally affected the patterns of S, Dmg, H, and J (Table 2).
Fig. 4

Changes in species number (S, a), richness (Dmg, b), diversity (H, c), and evenness (J, d) with grazing, rainfall manipulation, and seasonal variation. Different letters show significant variation in means (p < 0.05) among treatments. The error bars in the graph represent mean ± SD

The highest S (24.67 ± 0.57) and Dmg (5.38 ± 0.06) were observed in the G100% plot in January 2015 (Fig. 4a). The lowest H was in U50% (1.35 ± 0.26) and U150% (1.11 ± 0.17) plots in May 2016 compared to their counterparts in the grazed plots (Fig. 4c). J ranged from 0.50 ± 0.08 (U150%) to 0.82 ± 0.02 (G100%), with no significant mean differences among the grazed plots, unlike in the ungrazed plots where J reduced with 50% and 150% rainfall amounts.

The influence of grazing on herbaceous community was in the order: H (47.62%) > Dmg (45.06%) > S (35.89%) > J (32.09%).

Aboveground biomass (AGB)

The AGB destructively sampled was linearly correlated with their corresponding photographic images (pixels) in both grazed (r2 = 0.74–0.87; p < 0.05) and ungrazed (r2 = 0.83–0.91; p < 0.05) sites. AGB declined with grazing but increased with rainfall amount (Table 2). The G50% and U50% had the lowest AGB of 333.74 ± 5.27 g m−2 and 604.86 ± 89.48 g m−2 from grazed and ungrazed respectively, in September 2015 (Fig. 5a). The highest mean seasonal AGB was reported in the U100% plot, in May 2016 (Fig. 5a). However, for the entire measurement period, the U150% plot had more AGB of 1150 ± 119 g m−2 compared to the other plots (Fig. 5b). Mean AGB were different among the plots, except between G100 and G150% (p = 0.96). AGB was influenced by grazing, rainfall manipulation, seasonality in rainfall, and their interactions (Table 2). AGB sharply increased from the dry season (September) to the wet season (January), however, no significant difference was observed between the two wet seasons (January and May; Fig. 5a). H. fillipendula and B. decumbens constituted 66–75% of AGB from the ungrazed plots whereas P. dilatatum and B. insculpta contributed 26–36% of AGB in the grazed area (Fig. 5b).
Fig. 5

Seasonal changes in mean (plus SD) aboveground biomass (AGB) (a) showing variation between grazed and ungrazed plots. AGB increased significantly from the grazed to ungrazed along a rainfall gradient. The contribution of two most dominant species (b) from grazed; Paspalum dilatatum (PD), Bothriochloa insculpta (BI) and ungrazed; Brachiaria decumbens (BD) and Hyparrhenia fillipendula (HF) to AGB. Rainfall manipulation plots: 50, 100, and 150%

Relationships among parameters

The SEM used to predict the effect of VWC on herbaceous AGB and diversity had a good fit to the data (x2 = 2.98; p = 0.84; RMSEA < 0.05) (Fig. 6). However, we excluded the number of species (S) from the statistics, for the model to fit. Nevertheless, both Dmg and J indices still represented S (Table 1). The VWC had no direct effect on H and J but increased AGB and decreased Dmg (Fig. 6). On the contrary, the mediated (indirect-only mediation) effect of VWC on H and J through AGB and Dmg were significant (p < 0.05). Both H and J increased with Dmg but reduced significantly (p < 0.05) as AGB increased (Fig. 6). The changes in H were strongly correlated with J in our model (Fig. 6). Increased AGB significantly lowered H (r2 = 0.32; p < 0.05), particularly in the ungrazed plots (Fig. 7). The highest H was reported at an AGB range of 400–800 g m−2 in the grazed plot.
Fig. 6

Structural equation model (SEM) showing the effect of volumetric water content (VWC) on species diversity (H) and evenness (J) mediated by aboveground biomass (AGB) and the Margalef’s richness (Dmg). The Chi square value (X2 = 2.98; P = 0.84) expresses the model fit. The values presented alongside the arrows stand for the standardized regression coefficients. The e1–e4 attached to the rectangles are the error terms for every variable tested. The r2 values above the rectangles show the amount of variation explained by the model, while the thick arrows indicate the lack of significant effect where P > 0.05

Fig. 7

Linear regression showing relationship between aboveground biomass (AGB) and Shannon diversity index (H) plotted from grazed and ungrazed. The line is the significant regression fit (R2 = 0.32; P < 0.0001) across data points from all the plots sampled

Discussion

Patterns in soil water status

Changes in VWC (20 cm soil depth) in the plots were correlated with the amount of rain, an indication that rainwater was the main source of moisture input for soil and vegetation. On the other hand, livestock grazing modified the herbaceous structure and reduced the VWC through trampling and clipping of the vegetation. This exposed the soil surface, and accelerated soil moisture loss through evaporation, particularly during drought. Moreover, soil compaction by herbivores reduces the soil water holding capacity, hence, lowering the moisture levels in the grazed savannas (Holdo and Mack 2014). The interaction between grazing and variable rainfall input additionally accounts for the soil moisture variations in our plots (Table 2). Studies that support our findings show increased sensitivity of grazed ecosystems to changes in precipitation compared to their ungrazed counterparts (Frank 2007; Polley et al. 2008; Skinner et al. 2002). Such changes result from increased water deficit in the soil, prolonged drought (Frank 2007; Skinner et al. 2002), and effect of high radiation load (Polley et al. 2008).

Our findings revealed that, during drought, grazing had no effect on the changes in VWC at the 20 cm soil depth. This is shown by the lack of variation in VWC between G50 and U50% plots (Fig. 3a, b) in September 2015, contrary to our expectation. We anticipated that, the denser herbaceous canopy in the U50% plot compared to the G50% would aid in conserving soil moisture during drought. Conversely, the manipulative reduction of ambient rainfall dried up the soils, irrespective of whether the plots were grazed or not. According to Maestre et al. (2009), belowground competition by overlapping niches during drought diminishes the positive effects of dense plant canopies on soil moisture availability. Furthermore, modified microclimate conditions such as grazing exclusion to conserve soil water are less likely to remedy the effect of intense drought on the herbaceous layer community (Porensky et al. 2013). These findings partly explain the lack of variation in VWC between the G50 and U50% plots. We additionally suggest that the reduction of ambient rainfall during drought in the U50% plot triggered defoliation and plant death, making the soils vulnerable to water losses. Scholes and Archer (1997) in support of our results show that, drought imposed physiological stress causes premature leaf senescence and plant death, therefore, exposing the soils to higher evaporation rates due to reduced vegetation cover.

Studies from other ecosystems that conform to our findings, reveal that during drought, the soil moisture between grazed and ungrazed sites do not vary (K’Otuto et al. 2012; Sarmiento et al. 2004). According to Cingolani et al. (2005); Gu et al. (2006) and Skinner et al. (2002), the grazing history, functional group traits, variation in the Bowen ratio and root architecture of vegetation, largely account for soil moisture availability across ecosystems. Some belowground soil factors not tested by our study, such as niche differentiation and soil microbial activities (McCulley et al. 2007; Sankaran et al. 2004) explain further the soil moisture pattern between the G50 and U50% plots.

Changes in the herbaceous community composition and diversity

S, Dmg, H, and J declined in the rainfall manipulation plots, regardless of the decrease (50%) or increase (150%) in ambient rainfall (Fig. 4). We observed no effect of the interaction between rainfall manipulation and seasonality in rainfall on S, Dmg, H, and J. Consequently, our hypothesis that, the herbaceous species richness and diversity patterns positively correlate to rainfall gradient (Adler and Levine 2007; Cleland et al. 2013; Cornwell and Grubb 2003) was not supported. Nevertheless, livestock grazing interacted with rainfall manipulation and seasonality in rainfall to enhance S and Dmg in the herbaceous community. Grazing significantly (P < 0.05) influenced vegetation response to rainfall manipulation as depicted by the variation in Dmg, H and J (Table 2). Our results agree with Skinner et al. (2002) who observed that grazing modified species composition following supplemental summer precipitation. Anderson et al. (2007), on the other hand, showed that rainfall and soil phosphorous (P) strongly modulated the effect of grazing on the herbaceous species diversity. In our case, livestock grazing, influenced changes in the herbaceous community by regulating the distribution of biomass through changes in soil moisture.

In an Acacia-dominated savanna in Kenya, cattle grazing, in combination with periodic drought, enhanced plant diversity by creating open niches (micro-sites) for the cryptic species (Porensky et al. 2013). Their results partly explain the patterns of Dmg, H, and J reported in our grazed plots despite the reduced ambient rainfall or drought conditions in September (Fig. 4b–d). Gao et al. (2009), however, indicated that under short-term (2 year) exposures, the combined effect of grazing and drought lowered diversity and increased dominance compared to when such factors acted independently. This supports some of our observations. Our study did not compare the long and short-term effects of the interaction between grazing and rainfall variability on herbaceous diversity. Therefore, we were unable to determine the extent to which the herbaceous diversity shifted over long time.

The reduction of Dmg, H, and J in the 50% and 150% plots compared to the ambient in the ungrazed site (Fig. 4a–c) conforms to the findings of Anderson et al. (2007). Species richness declined at both the lower and higher rainfall sites along an annual precipitation gradient of 40–100 cm due to the shifting patterns of soil P following herbivore exclusion (Anderson et al. 2007). Consequently, grazing exclusion enhanced local extinction of species and reduced their colonization rates. We suggest that the lower diversity in U50% and U150% plots resulted from the inability of the rare species to compete for vital resources such as light and nutrients, and partly due to soil water stress, especially in the U50% plot. Porensky et al. (2013) report that, in the absence of grazing and drought, the more dominant perennial grasses and forbs out-compete the non-dominant species. Over time, grazing exclusion, potentially establishes a more stable, but, relatively homogenous herbaceous community (Anderson et al. 2007; Skinner et al. 2002).

The lower H, as observed in the ungrazed plots indicated the declining species number and evenness due to the absence of livestock, thus favoring the predominance of B. decumbens and H. fillipendula (Fig. 4b, c, Table 3). Livestock, by preferentially feeding on the dominant B. decumbens and H. fillipendula allows the establishment of other species, that would otherwise succumb to intense competition for resources. Preferential feeding on grasses by herbivores increases richness and evenness of forbs therefore modifying the herbaceous community structure (Augustine and Frank 2001; Koerner and Collins 2014). Despite resisting grazing pressure, the presence of B. insculpta and P. dilatatum as the most dominant species in the grazed plots did not suppress the less dominant species (Table 3). Consequently, this enhanced the coexistence amongst the herbaceous species, therefore, increasing their diversity. The stronger gap colonization potential and prostrate growth habit exhibited by the less dominant species (Gao et al. 2009), explain the pattern of diversity in our grazed plots. These strategies enhance herbaceous diversity by enabling species to withstand grazing pressure (Adler et al. 2001; Osborne et al. 2018). Numerous studies on plant community dynamics in the East Africa savanna (Anderson et al. 2007; Oba et al. 2001) and elsewhere (Bakker et al. 2003; Koerner and Collins 2014; Osem et al. 2002) agree with our findings.

Aboveground biomass dynamics (AGB)

Rainfall manipulation, grazing, seasonality in rainfall, and their interactions influenced the changes in AGB (Table 2). This is consistent with studies in the savannas that demonstrate the role of livestock grazing, rainfall, and its seasonality in regulating the herbaceous biomass (Anderson et al. 2007; Bat-Oyun et al. 2016; Metzger et al. 2005; Porensky et al. 2013). We attribute the significant drop in AGB in the grazed plots to clipping as earlier reported for the Lambwe Valley ecosystem (K’Otuto et al. 2012), in other parts of Kenya (Bat-Oyun et al. 2016; Kioko et al. 2012), and in savannas elsewhere (Dangal et al. 2016; Koerner and Collins 2014). AGB linearly responded to changes in VWC along a rainfall gradient of 50–150% (Fig. 5), an indication that the lower VWC from the grazed plots compared to the ungrazed contributed to the reduced AGB. Consequently, AGB decreased by 31.2% and increased by 8.47% due to the reduction and addition of ambient rain respectively. This implies that future reduction in rainfall would compromise sustainable biomass production.

Past studies which support our findings, conducted in the same area (K’Otuto et al. 2012; Otieno et al. 2010, 2011) showed that the annual AGB pattern strongly shifted with varying rainfall amounts. In our case, therefore, rainfall regulated AGB development, whereas grazing and seasonality in rainfall moderated the growth pattern of species that highly contributed to AGB. This was evident with a decline in AGB in September compared to January and May (Fig. 5) due to lower rainfall amount (Fig. 2). During this dry spell in September, mean AGB amongst the grazed plots (G50%, G100%, and G150%) were not significantly different due to intensified grazing on the lowly available vegetation. We expected increased AGB from G150% plot at this time compared to the other grazed plots (Fig. 5a) due to its higher VWC (Fig. 3b), but that was not the case. The G150% plot was most likely overgrazed due to its palatable foliage during drought resulting from conserved soil moisture.

Brachiaria decumbens and H. fillipendula contributed a greater proportion of AGB in the ungrazed plots, unlike in the grazed plots where livestock suppressed their dominance and allowed other species to flourish. Quattrocchi (2006) and Coughenour et al. (1985) reported that light grazing, high nitrogen concentration, and increased soil moisture promoted biomass development in B. decumbens and H. fillipendula. This explains the higher AGB of B. decumbens and H. fillipendula in our ungrazed plots, especially with 150% rainfall (Fig. 5b). Grasses dominated our ungrazed plots unlike the grazed ones where grasses and forbs coexisted.

The herbaceous community dynamics

Our model (Fig. 6) potentially explains the role of soil moisture as a regulator of the properties and functions of the herbaceous community in the Lambwe Valley savanna. Contrary to our expectations, there was no direct effect of VWC on H and J (Fig. 6). However, the mediated responses triggered by changes in AGB and Dmg explained the indirect influence of VWC on herbaceous diversity. By increasing AGB, both H and J declined, contrary to our expectation that higher AGB would enhance herbaceous species diversity due to improved soil moisture availability. The reductive influence of VWC on Dmg, however, increased H and J. The strong effect of livestock grazing on our sites, and other confounding factors, not tested in our study, explain these patterns. The lower VWC and AGB in our grazed plots hindered the establishment of predominant species, therefore increasing Dmg, H, and J (Figs. 3, 4 and 5), which enhanced the herbaceous species diversity.

The negative correlation between VWC and Dmg, and the lack of direct effect of the former on H and J (Fig. 6) was surprising. Knapp et al. (2002) reported increased species diversity along a soil moisture gradient in a mesic grassland. Koerner and Collins (2014) however, showed that, acting independently, drought minimally influenced the plant community structure due to the stronger effect of grazing. Similarly, higher rainfall in a savanna ecosystem reduced species diversity compared to lower rainfall amount (Barbosa da Silva et al. 2016). These findings partly explain the trend in correlation between VWC and Dmg, H and J especially, under manipulated rainfall amount. The positive correlation between VWC and AGB (Fig. 6) shows the importance of soil moisture as a major driver of herbaceous biomass development. Ecosystems with similar attributes like those of Lambwe Valley report higher AGB with increasing soil moisture (Augustine and McNaughton 2006; Frank 2005).

The number of species, abundance and evenness in their distribution were significant in driving herbaceous species diversity (H). This was demonstrated by the stronger correlations amongst Dmg, H, and J (Fig. 6), emphasizing the role of grazing (Table 2) in modifying the herbaceous community structure by reducing the dominance of B. insculpta and P. dilatatum. The lower H (Fig. 4) in the ungrazed plots resulted from the dominance of B. decumbens and H. fillipendula. These species imposed competitive advantage in resource utilization over the less dominant species, mostly forbs, hence their higher proportion of AGB (Fig. 5b). Empirical studies that support our results show that dominance thrives on greater competitive ability of species in resource utilization, whereas, varying grazing intensities facilitate species diversity and richness (Adler et al. 2011; Augustine and Frank 2001; Bakker et al. 2003). However, these explanations only apply to moderately grazed ecosystems, since intensive grazing hinders species dominance due to selective clipping (Fraser et al. 2013; Graham and Duda 2011; Tilman et al. 2012). Higher diversity occurred in the grazed plots, characterized by lower AGB (Fig. 7), suggesting that livestock grazing moderated the herbaceous community structure. Findings from grass-dominated ecosystems (Loreau et al. 2001; Schläpfer and Schmid 1999) reveal positive correlations between aboveground biomass and species diversity due to facilitation and niche complementarity. In the Kruger National Park in South Africa, the herbaceous species diversity declined with increasing biomass in the absence of grazing (van Coller et al. 2013), agreeing with our current findings. The reduction in biomass and canopy cover due to grazing promotes spatial heterogeneity in resources, enhancing the coexistence of herbaceous species and species diversity (Bakker et al. 2003). Studies conducted in East Africa and other sub-humid and semi-arid ecosystems elsewhere demonstrate the positive role of moderate grazing on the herbaceous species diversity (Anderson et al. 2007; Dingaan et al. 2016; Koerner and Collins 2014; Oba et al. 2001; Osem et al. 2002).

Conclusion

Our results demonstrate that rainfall amount and its seasonality are key determinants of changes in the herbaceous community structure with grazing as a modifier. The effect of grazing on vegetation enhances herbaceous species diversity through reduced competition for resources by the dominant species: B. decumbens and H. fillipendula. We illustrate that ambient rainfall favors higher diversity, with a sharp decline in the latter resulting from reduced or increased rainfall amounts. From our findings, the aboveground biomass positively correlated with soil moisture, a function of rainfall input. However, aboveground biomass outside the 400–800 gm−2 range lowered the herbaceous species diversity.

We project lower herbaceous species diversity and aboveground biomass in the Lambwe Valley ecosystem following decreased rainfall from the current amount. By enhancing evenness through facilitated coexistence of species, livestock grazing will increase species diversity at ambient rainfall. The interaction between grazing and rainfall manipulation will drive the changes in diversity and aboveground biomass. The mediated responses of the aboveground biomass and species composition to varying soil moisture, will additionally enhance the herbaceous species diversity. This research provides a basis for future studies on the implications of livestock grazing and rainfall variability on sustainable management of savanna ecosystems.

Notes

Acknowledgements

This research was funded by the British Ecological Society (BES), International Foundation for Science (IFS) and National Council for Science Technology and Innovation (NACOSTI)-Kenya. We are grateful for their support in logistics and procurement of equipment.

Compliance with ethical standards

Conflict of interest

The authors declare that they have no conflict of interest.

References

  1. Adler PB, Levine JM (2007) Contrasting relationships between precipitation and species richness in space and time. Oikos 116:221–232CrossRefGoogle Scholar
  2. Adler PB, Raff D, Lauenroth W (2001) The effect of grazing on the spatial heterogeneity of vegetation. Oecologia 128:465–479CrossRefPubMedGoogle Scholar
  3. Adler PB, Seabloom EW, Borer ET et al (2011) Productivity is a poor predictor of plant species richness. Science 333:1750CrossRefPubMedGoogle Scholar
  4. Alley R, Marotzke T, Nordhaus W et al (2003) Abrupt climate change. Science 299:2005–2010CrossRefPubMedGoogle Scholar
  5. Allsopp R, Baldry DAT (1972) A general description of the Lambwe Valley area of South Nyanza District, Kenya. B World Health Organ 47:691–697Google Scholar
  6. Anderson TM, Ritchie ME, McNaughton SJ (2007) Rainfall and soils modify plant community response to grazing in Serengeti National Park. Ecology 88:1191–1201CrossRefGoogle Scholar
  7. Arbuckle JL (2017) Amos 25.0: computer program. IBM SPSS, ChicagoGoogle Scholar
  8. Augustine DJ (2003) Spatial heterogeneity in herbaceous layer of a semi-arid savanna ecosystem. Plant Ecol 167:319–332CrossRefGoogle Scholar
  9. Augustine DJ, Frank DA (2001) Effects of migratory grazers on spatial heterogeneity of soil nitrogen properties in a grassland ecosystem. Ecology 82:3149–3162CrossRefGoogle Scholar
  10. Augustine DJ, McNaughton SJ (2006) Interactive effects of ungulate herbivores, soil fertility and variable rainfall on ecosystem processes in a semi-arid savanna. Ecosystems 9:1242–1256CrossRefGoogle Scholar
  11. Bakker C, Blair JM, Knapp AK (2003) Does resource availability, resource heterogeneity or species turnover mediate changes in plant species richness in grazed grasslands? Oecologia 137:385–391CrossRefPubMedGoogle Scholar
  12. Barbosa da Silva FH, Arieira J, Parolin P, Nunes da Cunha C, Junk WJ (2016) Shrub encroachment influences herbaceous communities in flooded grasslands of a neotropical savanna wetland. Appl Veg Sci 19:391–400CrossRefGoogle Scholar
  13. Bardgett RD, Wardle DA (2003) Herbivore-mediated linkages between above ground and below ground communities. Ecology 84:2258–2268CrossRefGoogle Scholar
  14. Bat-Oyun T, Shinoda M, Cheng Y, Purevdorj Y (2016) Effects of grazing and precipitation variability on vegetation dynamics in a Mongolian dry steppe. J Plant Ecol 9:508–519CrossRefGoogle Scholar
  15. Beier C, Beierkuhnlein C, Wohlgemuth T et al (2012) Precipitation manipulation experiments—challenges and recommendations for the future. Ecol Let 15:899–911CrossRefGoogle Scholar
  16. Bonham CD (1989) Measurements for terrestrial vegetation. Wiley, New YorkGoogle Scholar
  17. Borer ET, Seabloom EW, Gruner DS et al (2014) Herbivores and nutrients control grassland plant diversity via light limitation. Nature 508:517–520CrossRefGoogle Scholar
  18. Case M (2006) Climate change impacts on East Africa: a review of scientific literature. Worldwide Fund for Nature, SwitzerlandGoogle Scholar
  19. Cingolani AM, Noy-Meir I, Diaz S (2005) Grazing effects on rangeland diversity: a synthesis of contemporary models. Ecol Appl 15:757–773CrossRefGoogle Scholar
  20. Cleland EE, Collins SL, Dickson TL et al (2013) Sensitivity of grassland plant community composition to spatial vs. temporal variation in precipitation. Ecology 94:1687–1696CrossRefPubMedGoogle Scholar
  21. Conway G, (2009) The science of climate change in Africa: impacts and adaptation. Discussion paper No. 1. Imperial College London, Grantham Institute for Climate Change, LondonGoogle Scholar
  22. Cornwell WK, Grubb PJ (2003) Regional and local patterns in plant species richness with respect to resource availability. Oikos 100:417–428CrossRefGoogle Scholar
  23. Coughenour MB, Mc Naughton SJ, Wallace CC (1985) Response of an African tall grass (Hyparrhenia filipendula Stapf.) to defoliation and limitation of water and nitrogen. Oecologia 68:80–86CrossRefGoogle Scholar
  24. Cumming DHM (1982) The influence of large herbivores on savanna structure in Africa. In: Huntley BJ, Walker BH (eds) Ecology of tropical savannas. Ecol stud, vol 42. Springer-Verlag, Berlin, pp 217–245CrossRefGoogle Scholar
  25. Dangal SRS, Tian H, Lu C, Pan S, Pederson N, Hessl A (2016) Synergistic effects of climate change and grazing on net primary production of Mongolian grasslands. Ecosphere 7:e01274.  https://doi.org/10.1002/ecs2.1274 CrossRefGoogle Scholar
  26. Davidson O, Halsnaes K, Huq S et al (2003) The development and climate nexus: the case of sub-Saharan Africa. Clim Policy 3:97–113CrossRefGoogle Scholar
  27. Dingaan MNV, Walker S, Tsubo M, Newby T (2016) Influence of grazing on plant diversity productivity relationship in semi-arid grassland of S. Africa. Appl Ecol Env Res 14:1–13CrossRefGoogle Scholar
  28. Frank DA (2005) The interactive effects of grazing ungulates and aboveground production of grassland diversity. Oecologia 143:629–634CrossRefPubMedGoogle Scholar
  29. Frank DA (2007) Drought effects on above and below ground production of a grazed temperate grass-land ecosystem. Oecologia 152:131–139CrossRefGoogle Scholar
  30. Fraser LH, Henry HA, Carlyle CN et al (2013) Coordinated distributed experiments: an emerging tool for testing global hypotheses in ecology and environmental science. Front Ecol Environ 11:147–155CrossRefGoogle Scholar
  31. Gao YZ, Giese M, Han XG et al (2009) Land use and drought interactively affect interspecific competition and species diversity at the local scale in a semiarid steppe ecosystem. Ecol Res 24:627–635CrossRefGoogle Scholar
  32. Grace JB, Anderson TM, Olff H, Scheiner SM (2010) On the specification of structural equation models for ecological systems. Ecol Monogr 80:67–87CrossRefGoogle Scholar
  33. Graham JH, Duda JJ (2011) The humpbacked species richness-curve: a contingent rule for community ecology. Int J Ecol 11:1–16CrossRefGoogle Scholar
  34. Gu L, Meyers T, Pallardy SG et al (2006) Direct and indirect effects of atmospheric conditions and soil moisture on surface energy partitioning revealed by a prolonged drought at a temperate forest site. J Geophys Res 15:6.  https://doi.org/10.1029/s2006jd007161 CrossRefGoogle Scholar
  35. Hanke W, Böhner J, Dreber N et al (2014) The impact of livestock grazing on plant diversity: an analysis across dryland ecosystems and scales in southern Africa. Ecol Appl 245:1188–1203CrossRefGoogle Scholar
  36. Harrison KA, Bardgett RD (2004) Browsing by red deer negatively impacts on soil nitrogen availability in regenerating woodland. Soil Biol Biochem 36:115–126CrossRefGoogle Scholar
  37. Holdo RM, Mack MC (2014) Functional attributes of savanna soils: contrasting effects of tree canopies and herbivores on bulk density, nutrients, and moisture dynamics. J Ecol 102:1171–1182CrossRefGoogle Scholar
  38. Hulme M, Doherty R, Ngara T, New M, Lister D (2001) African climate change: 1900-2100. Clim Res 17:145–168CrossRefGoogle Scholar
  39. Jones SK, Collins SL, Blair JM, Smith MD, Knapp AK (2016) Altered rainfall patterns increase forb abundance and richness in native tallgrass prairie. Sci Rep.  https://doi.org/10.1038/srep20120 CrossRefPubMedPubMedCentralGoogle Scholar
  40. K’Otuto GO, Otieno DO, Seo B, Ogindo HO, Onyango JC (2012) Carbon dioxide exchange and biomass productivity of herbaceous layer of a managed tropical humid savanna ecosystem in Western Kenya. J Plant Ecol 6:286–297CrossRefGoogle Scholar
  41. K’Otuto GO, Otieno DO, Onyango JC, Ogindo HO (2014) Seasonal dynamics in carbon dioxide fluxes of the herbaceous layer of a moist Kenyan savannah. Appl Ecol Environ Res 12:63–82CrossRefGoogle Scholar
  42. Kioko J, Kiringe JW, Ole Seno S (2012) Impacts of livestock on grazing on a savanna grassland in Kenya. J Arid Land 4:29–35CrossRefGoogle Scholar
  43. Knapp AK, Fay PA, Blair JM et al (2002) Rainfall variability, carbon cycling and plant species diversity in a mesic grassland. Science 298:2202–2205CrossRefGoogle Scholar
  44. Knapp A, Beier C, Briske D et al (2008) Consequences of altered precipitation regimes for terrestrial ecosystems. Bioscience 58:1–11CrossRefGoogle Scholar
  45. Koerner SE, Collins SL (2014) Interactive effect of grazing, drought, and fire on grassland plant communities in Northern America and South Africa. Ecology 95:98–109CrossRefGoogle Scholar
  46. Loreau M, Naeem S, Inchausti P et al (2001) Biodiversity and ecosystem functioning: current knowledge and future challenges. Science 294:804–808CrossRefGoogle Scholar
  47. Maestre FT, Callaway RM, Valladares F, Lortie CJ (2009) Refining the stress-gradient hypothesis for competition and facilitation in plant communities. J Ecol 97:199–205CrossRefGoogle Scholar
  48. Magurran AE (2004) Measuring biological diversity, 2nd edn. Blackwell Science Ltd, OxfordGoogle Scholar
  49. Marshall F, Reid REB, Goldstein S et al (2018) Ancient herders enriched and restructured African grasslands. Nature 561:387–390CrossRefPubMedGoogle Scholar
  50. McCulley RL, Boutton TW, Archer SR (2007) Soil respiration in a Subtropical Savanna Parkland: response to water additions. Soil Sci Soc Am J 71:820–828CrossRefGoogle Scholar
  51. Merbold L, Ardö J, Arneth A et al (2009) Precipitation as driver of carbon fluxes in 11 African ecosystems. Biogeosciences 6:1027–1041CrossRefGoogle Scholar
  52. Metzger KL, Coughenour MB, Reich RM, Boone RB (2005) Effects of seasonal grazing on plant species diversity and vegetation structure in a semi-arid ecosystem. J Arid Environ 61:147–160CrossRefGoogle Scholar
  53. Milchunas DG, Sala OE, Lauenroth WK (1988) A generalized model of the effects of grazing by large herbivores on grassland community structure. Am Nat 132:87–106CrossRefGoogle Scholar
  54. Mouillot D, Graham NAJ, Villéger S et al (2013) A functional approach reveals community responses to disturbances. Trends Ecol Evol 28:167–177CrossRefGoogle Scholar
  55. Nippert JB, Knapp AK, Briggs JM (2006) Intra-annual rainfall variability and grassland productivity: can the past predict the future? Plant Ecol 184:65–74CrossRefGoogle Scholar
  56. Oba G, Vetaas OR, Stenseth NC (2001) Relationships between biomass and plant species richness in arid zone grazing lands. J Appl Ecol 38:345–836CrossRefGoogle Scholar
  57. Olff H, Ritchie ME (1998) Effects of herbivores on grassland plant diversity. Trends Ecol Evol 13:261–265CrossRefPubMedGoogle Scholar
  58. Osborne CP, Charles-Dominique T, Stevens N et al (2018) Human impacts in African savannas are mediated by plant functional traits. New Phytol 220:10–24CrossRefPubMedGoogle Scholar
  59. Osem Y, Perevolotsky A, Kigel J (2002) Grazing effect on diversity of annual plant communities in a semi-arid rangeland: interactions with small-scale spatial and temporal variation in primary productivity. J Ecol 90:936–946CrossRefGoogle Scholar
  60. Otieno DO, K’Otuto GO, Maina JN, Kuzyakov Y, Onyango JC (2010) Responses of ecosystem carbon dioxide fluxes to soil moisture fluctuations in a moist Kenyan savanna. J Trop Ecol 26:605–618CrossRefGoogle Scholar
  61. Otieno DO, K’Otuto GO, Jakli B, Maina JN, Jung E, Onyango JC (2011) Spatial heterogeneity in ecosystem structure and productivity in a moist Kenyan Savanna. J Ecol 212:769–783Google Scholar
  62. Polley HW, Frank AB, Sanabria J, Phillips RL (2008) Interannual variability in carbon dioxide fluxes and flux-climate relationships on grazed and ungrazed northern mixed-grass prairie. Glob Change Biol 14:1620–1632CrossRefGoogle Scholar
  63. Porensky LS, Wittman C, Riginos C, Young T (2013) Herbivory and drought interact to enhance spatial patterning and diversity in a savanna understory. Oecologia 173:591–602CrossRefPubMedGoogle Scholar
  64. Preacher KJ, Hayes AF (2008) Asymptotic and resampling strategies for assessing and comparing indirect effects in multiple mediator models. Behav Res Methods 40:879–891CrossRefPubMedGoogle Scholar
  65. Quattrocchi U (2006) CRC World Dictionary of grasses: common names, scientific names, eponyms and etymology, vol 3. CRC Press, LondonCrossRefGoogle Scholar
  66. Sankaran M, Ratmann J, Hanan NP (2004) Tree grass co-existence in savannas revisited. Insights from an examination of assumptions and mechanisms involved in existing models. Ecol Lett 7:480–490CrossRefGoogle Scholar
  67. Sarmiento G, Pinillos M, Pereira da Silva M, Acevedo D (2004) Effects of soil water regime and grazing on vegetation diversity and production in a hyperseasonal savanna in the Apure Llanos, Venezuela. J Trop Ecol 20:209–220CrossRefGoogle Scholar
  68. Savadogo P, Tiveau D, Sawadogo L, Tigabu M (2008) Herbaceous species responses to long-term effects of prescribed fire, grazing and selective tree cutting in the savanna-woodlands of West Africa. Perspect Plant Ecol Evol Syst 10:179–195CrossRefGoogle Scholar
  69. Schläpfer F, Schmid B (1999) Ecosystem effects of biodiversity: a classification of hypotheses and exploration of empirical results. Ecol Appl 9:893–912CrossRefGoogle Scholar
  70. Schneider CA, Rasband WS, Eliceiri KW (2012) NIH image to image J: 25 years of image analysis. Nat Methods 9:671–675CrossRefPubMedPubMedCentralGoogle Scholar
  71. Scholes RJ, Archer SR (1997) Tree-grass interaction in savanna. Annu Rev Ecol Syst 28:517–544CrossRefGoogle Scholar
  72. Skinner RH, Hanson JD, Hutchinson GL, Shuman GE (2002) Response of C3 and C4 grasses to supplemental summer precipitation. J Range Manag 55:517–522CrossRefGoogle Scholar
  73. Swemmer AM, Knapp AK, Snyman HA (2007) Intra-seasonal precipitation patterns and aboveground productivity in three perennial grasslands. J Ecol 95:780–788CrossRefGoogle Scholar
  74. Tackenberg O (2007) A new method for non-destructive measurement of biomass, growth rates, vertical biomass distribution and dry matter content based on digital image and analysis. Ann Bot 99:777–783CrossRefPubMedPubMedCentralGoogle Scholar
  75. Tilman D, Reich PB, Isbel F (2012) Biodiversity impacts ecosystem productivity as much as resources disturbance or herbivory. PNAS 109:10394–10397CrossRefPubMedGoogle Scholar
  76. van Coller H, Siebert F, Siebert SJ (2013) Herbaceous species diversity patterns across various treatments of herbivory and fire along the sodic zone of the Nkuhlu exclosures, Kruger National Park. Koedoe.  https://doi.org/10.4102/koedoe.v55i1.1112 CrossRefGoogle Scholar
  77. van der Waal C, Kool A, Meijer SS (2011) Large herbivores may alter vegetation structure of semi-arid savannas through soil nutrient mediation. Oecologia 165:1095–1107CrossRefPubMedPubMedCentralGoogle Scholar
  78. William CA, Albertson TD (2004) Soil moisture control on canopy scale water and carbon fluxes in an African Savanna. Water Resour Res 40:1–14Google Scholar
  79. Zeppel MJB, Wilks J, Lewis J (2014) Impacts of extreme precipitation and seasonal changes in precipitation on plants. Biogeosciences 11:3083–3093CrossRefGoogle Scholar

Copyright information

© The Botanical Society of Japan and Springer Japan KK, part of Springer Nature 2019

Authors and Affiliations

  • Daniel Osieko Okach
    • 1
    Email author
  • Joseph O. Ondier
    • 2
  • Gerhard Rambold
    • 3
  • John Tenhunen
    • 1
  • Bernd Huwe
    • 4
  • Eun Young Jung
    • 1
  • Dennis O. Otieno
    • 1
    • 5
  1. 1.Department of Plant EcologyUniversity of BayreuthBayreuthGermany
  2. 2.Department of BotanyMaseno UniversityMasenoKenya
  3. 3.Department of MycologyUniversity of BayreuthBayreuthGermany
  4. 4.Department of Soil PhysicsUniversity of BayreuthBayreuthGermany
  5. 5.Jaramogi Oginga Odinga University of Science and TechnologyBondoKenya

Personalised recommendations