Effect of grazing pressure on the plant community structure and composition in Castril, Santiago and Pontones pastoral commons
Framework
In pastoral ecology, we can employ grazing indicators as proxies of the actual grazing pressure. These indicators can concern different spatial levels of a pastoral system. At the field level, indicators can be a measure of local grazing pressure. In contrast, stocking rate is grazing indicator at the individual grazing area level (I employed the level terms of hierarchical structure of a pastoral system of Balent and Stafford-Smith 1991).
In this framework we could wonder:
Which are the relationship between grazing indicators ?
To what extent the vegetation variability is related to the grazing indicators ?
Are there significant differences of grazing indicators across commons and transhumance modalities ?
Which are the relationship between plant community types* of CSP and the grazing indicators ?
*plant community types found in the paper that we already submitted.
In this report I focus on research questions 1 and 2 and propose a statistical analysis for question 3 and 4 that we could discuss.
Materials and methods
Field level grazing indicators
Indicators measured at the field level, i.e.nearby or along the 72 vegetation transects carried out during spring 2022 and 2023.
Plant Utilization Rates (PUR): We followed the method described by Ruiz-Mirazo et al. (2011) which proposes qualitative categories of plant utilization ranging from 0 to 5, with 3 referring to intermediate utilization. PUR have been measured nearby the vegetation transect in 25 individuals of dominant species, i.e. Helianthemum cinereum, Helianthemum oelandicum, Helianthemum appenninum, Thymus serpylloides, Koeleria vallesiana, Seseli montanum and Festuca segimonensis). Also we measured an overall Plant Utilisation Rate (PUR) of the vegetation
pur_RA, which was measured at each meter of the 10m transect. For the statistical analysis we aggregated species with similar PUR. We generated two categories;low palatablespecies (PUR < 2) including Koeleria vallesiana, Seseli montanum and Festuca segimonensis, andhigh palatablespecies (PUR >4) including species of Helianthemum spp. and Thymus serpylloides.pl_density: total number of plant contacts in the transects. Plant density allows evaluating grazing effects (Castañeda et al., 2023).n_flower: total number of plant with flowers recorded in the transect. Flower abundance has been negatively associated with grazing pressure (Tadey et al. 2015).bare: Bare ground recorded in the transect. Bare ground can be an indicator of herbivores’ activity and is positively correlated with runoff and soil erosion (Pyke et al. 2002).dung: Dung measured at the beginning (dung_sp) and at the end (dung_au) of the grazing period. We employed quadrats of 50 x 50 cm to measure the dung amount. The quadrats were located adjacent to the 10m transect every 1 meter so carrying out 10 dung amount measurement. Dung can be a measure of grazing pressure on the transect scale (Jordan et al. 2022).
Individual grazing area level
Stocking Rate,which refers to the number of livestock per hectare over a specified time (Allen et al. 2011), is usually used to asses the effect of livestock on vegetation. Here we used animals*days per hectare (Scarnecchia 1985), which consider the number of days that the comarca is grazed. We gather such informations through the observations in field of Pau and Francisco and our observations during 2022 and 2023 fieldwork. Also, I check the transcriptions of the interviews done by me and Adrià during 2023.
\[\text{Stocking Rate} = \frac{Animals * Days}{Hectares}\]
However, this indicator may not be suitable in case of environmental and plant community heterogeneity, also it does not take into account herding practices at the local scale (Genin and Hanafi 2010).
Data summary by indicator
## stoking_rate_2 pl_density pur_RA pur_high_palatable
## Min. : 199.7 Min. : 93.0 Min. :3.200 Min. :4.12
## 1st Qu.: 371.4 1st Qu.:123.0 1st Qu.:3.900 1st Qu.:4.64
## Median : 483.5 Median :139.5 Median :4.200 Median :4.75
## Mean : 487.6 Mean :148.0 Mean :4.193 Mean :4.75
## 3rd Qu.: 592.5 3rd Qu.:172.2 3rd Qu.:4.525 3rd Qu.:4.92
## Max. :1035.8 Max. :297.0 Max. :4.900 Max. :5.00
## pur_low_palatable dung_au dung_sp n_flower
## Min. :0.40 Min. : 22.00 Min. : 11.00 Min. : 3.00
## 1st Qu.:1.00 1st Qu.: 57.75 1st Qu.: 54.00 1st Qu.:18.00
## Median :1.49 Median : 96.00 Median : 81.00 Median :24.00
## Mean :1.52 Mean :106.60 Mean : 96.76 Mean :25.11
## 3rd Qu.:1.96 3rd Qu.:146.50 3rd Qu.:117.00 3rd Qu.:31.25
## Max. :3.04 Max. :300.00 Max. :392.00 Max. :69.00
## bare
## Min. : 0.000
## 1st Qu.: 4.000
## Median : 7.000
## Mean : 7.611
## 3rd Qu.:11.250
## Max. :17.000First results
1. Which are the relationship between grazing indicators ?
Spearman correlation between indicators.
Correlation matrix:
## stoking_rate_2 pl_density pur_RA pur_high_palatable
## stoking_rate_2 1
## pl_density -0.12 1
## pur_RA 0.27 -0.16 1
## pur_high_palatable 0.25 0.19 0.37 1
## pur_low_palatable 0.06 -0.27 0.26 -0.51
## dung_au 0.21 -0.02 0 0.17
## dung_sp 0.01 -0.09 0.16 0.06
## n_flower -0.1 0.17 -0.08 -0.18
## bare 0 -0.53 0.12 -0.2
## pur_low_palatable dung_au dung_sp n_flower bare
## stoking_rate_2
## pl_density
## pur_RA
## pur_high_palatable
## pur_low_palatable 1
## dung_au -0.25 1
## dung_sp 0.12 0.42 1
## n_flower 0.03 -0.22 -0.05 1
## bare 0.25 -0.13 0.03 -0.12 1The indicators are not highly correlated between them. As they are not highly correlated, we do not have the issue of multicollinearity so we can keep them all to the PCA analysis.
PCA of grazing indicators
## $importance
## Importance of components:
## PC1 PC2 PC3 PC4 PC5 PC6 PC7
## Eigenvalue 1.9521 1.8546 1.3366 1.1694 0.82805 0.67062 0.5301
## Proportion Explained 0.2169 0.2061 0.1485 0.1299 0.09201 0.07451 0.0589
## Cumulative Proportion 0.2169 0.4230 0.5715 0.7014 0.79341 0.86792 0.9268
## PC8 PC9
## Eigenvalue 0.44940 0.20923
## Proportion Explained 0.04993 0.02325
## Cumulative Proportion 0.97675 1.00000Principal component one explains 22%, component two explains 21%, and component three explains 15% of the variance.
inv_scores = data.frame(ind.pca$CA$u)
inv_scores1 <- right_join(rownames_to_column(df_grazing_ind_v2), rownames_to_column(data.frame(inv_scores)), by = "rowname")
# add categories like common instead of df_grazing_ind_v2 and this code has sense.
var_scores = data.frame(ind.pca$CA$v)
Variables and principal components 1-2
Transects and principal Components 1-2
2. To what extent the variability of the dry perennial herbaceous community is related to the grazing indicators ?
distance-based Redundancy Analysis (db-RDA): Modelling of grazing indicators and vegetation data
# capscale() with raw (site by species) data (Borcard et al. 2018)
rda_22_23_2_bis <- capscale(community_data_density_22_23.72_sin_cus ~ stoking_rate_2 + pl_density + pur_RA + pur_high_palatable + pur_low_palatable + dung_au + dung_sp + n_flower + bare + Condition(scores(as.matrix(coord_22_23))), data = df_grazing_ind_v2, distance = "bray",add = "lingoes")
rda_22_23_2_bis_sum = summary(rda_22_23_2_bis)
# Total variance explained by the model
(Tot.var <- rda_22_23_2_bis$tot.chi)## [1] 23.46159# all eigenvalues
eig.val <- c(rda_22_23_2_bis$CCA$eig, rda_22_23_2_bis$CA$eig)
# Relative eigenvalues of Y-hat
eig.val.rel <- eig.val / Tot.var
# Varibility explained by the axes 1,2 and 3
100*eig.val.rel[1:3]## CAP1 CAP2 CAP3
## 5.780011 4.188453 2.772724##
## -----------------------------------------------------------------------
## Site constraints (lc) selected. To obtain site scores that are weighted
## sums of species scores (default in vegan), argument site.sc must be set
## to wa.
## -----------------------------------------------------------------------
Figure: Grazing indicators and vegetation transects. Angles between variables reflect their correlation
##
## -----------------------------------------------------------------------
## Site constraints (lc) selected. To obtain site scores that are weighted
## sums of species scores (default in vegan), argument site.sc must be set
## to wa.
## -----------------------------------------------------------------------
Figure: Grazing indicators and species. Overlapped species positively associated with axis 2 are Xeranthemum inapertum and Rumex bucephalophorus.
Even though the variability explained by our model is low, we figure out interesting patterns regarding species repartition and grazing indicators. The negative size of axis 1 holds transects having high bare ground, dung amount during the spring and high PUR of low palatable species. In contrast, positive size of axis 1 holds transects having high plant density, yet also and surprisingly, high PUR of high palatable species. Festuca segimomensis and Poa bulbosa are positively related to axis 1.
Concerning axis 2, the overall PUR of vegetation
(pur_RA) is positively related to axis 2 and also the PUR
of high and low palatable specie. Annuals plants, i.e. Xeranthemum
inapertum,Rumex bucephalophorus and Aegilops
spp., are positively related to axis 2. On the other hand, flower
abundance and plant density are negatively related to axis 2.
Perennials, i.e.Helianthemum spp., Thymus serpylloides
and Festuca segimomensis are negatively associated with axis
2.
Stoking rate is not related to axis 1 and 2. So, it seems that the stocking rate is not behind the vegetation variability.
Research Questions (RS) 3 and 4
Considering that we have a group classification of transects (e.g. transhumance modalities for RS.3 or plant community types for RS.4) based on the vegetationdata and we try to explain this classification using our grazing indicators, we could perform a Linear Discriminant Analysis (LDA).
References
Allen, V. G., Batello, C., Berretta, E. J., Hodgson, J., Kothmann, M., Li, X., McIvor, J., Milne, J., Morris, C., Peeters, A., Sanderson, M., & The Forage and Grazing Terminology Committee. (2011). An international terminology for grazing lands and grazing animals. Grass and Forage Science, 66(1), 2–28. https://doi.org/10.1111/j.1365-2494.2010.00780.x
Balent, G., & Stafford-Smith, M. (1991). Conceptual model for evaluating the consequences of management practices on the use of pastoral resources. Proceedings from the Fourth International Rangeland Congress, 1158–1164.
Castañeda, I., Callède, L., & Corcket, E. (2023). Apports d’un dispositif intégré « Communauté-Population-Sol » pour évaluer l’effet des grands herbivores sur les écosystèmes. Naturae, 3. https://doi.org/10.5852/naturae2023a3
Genin, D., & Hanafi, A. (2010). Estimating pastoral pressure on arid rangelands: How to take into account herd management practices? Livestock Research for Rural Development, 22(4). http://www.lrrd.org/lrrd22/4/geni22078.htm
Jordan, S. E., Palmquist, K. A., Burke, I. C., & Lauenroth, W. K. (2022). Small effects of livestock grazing intensification on diversity, abundance, and composition in a dryland plant community. Ecological Applications, 32(8), e2693. https://doi.org/10.1002/eap.2693
Pyke, D. A., Herrick, J. E., Shaver, P., & Pellant, M. (2002). Rangeland Health Attributes and Indicators for Qualitative Assessment. Journal Range Management, 15.
Scarnecchia, D. L. (1985). The Relationship of Stocking Intensity and Stocking Pressure to Other Stocking Variables. Journal of Range Management, 38(6), 558. https://doi.org/10.2307/3899753
Ruiz-Mirazo, J., Robles, A. B., & González-Rebollar, J. L. (2011). Two-year evaluation of fuelbreaks grazed by livestock in the wildfire prevention program in Andalusia (Spain). Agriculture, Ecosystems & Environment, 141(1–2), 13–22. https://doi.org/10.1016/j.agee.2011.02.002
Tadey, M. (2015). Indirect effects of grazing intensity on pollinators and floral visitation. Ecological Entomology, 40(4), 451–460. https://doi.org/10.1111/een.12209