grazing index

The grazing index we proposed was calculated by comparing photosynthetic biomass before (NDVI07, which means NDVI calculated in July) and after (NDVIL06, which means NDVI calculated in June) grazing disturbance.

(It is important to note that the grazing season begins in June on our study sites)

load data

rm(list=ls(all=TRUE))
shrubs_df <- readRDS("shrubs_dataframe.RDS")

Grazing index was calculated by subtracting the NDVI in June (little or no visible impact of grazing on biomass) from the NDVI in July (grazing disturbance on vegetation - avoiding August water restriction).

Positive values indicate higher disturbance.

shrubs_df$Grz07L06 <-  shrubs_df$ndviL06 - shrubs_df$ndvi07

For the purpose of validating the grazing index, we tested the correlation between the grazing index and grazing proxy data gathered from the field.

cor.test(shrubs_df$Grz07L06, shrubs_df$trump, method=c("pearson"))
## 
##  Pearson's product-moment correlation
## 
## data:  shrubs_df$Grz07L06 and shrubs_df$trump
## t = 2.2451, df = 41, p-value = 0.03022
## alternative hypothesis: true correlation is not equal to 0
## 95 percent confidence interval:
##  0.03390437 0.57416383
## sample estimates:
##       cor 
## 0.3308789
cor.test(shrubs_df$Grz07L06, shrubs_df$bite, method=c("pearson"))
## 
##  Pearson's product-moment correlation
## 
## data:  shrubs_df$Grz07L06 and shrubs_df$bite
## t = 2.3544, df = 41, p-value = 0.02342
## alternative hypothesis: true correlation is not equal to 0
## 95 percent confidence interval:
##  0.04993939 0.58483251
## sample estimates:
##      cor 
## 0.345107

grazing abandonment

To understand the pattern of abandonment of grazed pastures, I computed the difference between the grazing index in the present (2012) and the past (1985)

Negative values indicate grazing decrease

shrubs_df$var_Grz07L06 <- ( (shrubs_df$Grz07L06) - (shrubs_df$past_Grz07E06) )

#correlation between grazing abandonment and current grazing pattern

Then, I wish to know whether the low-grazed areas have previously been grazed and, therefore, have been abandoned by the grazing

cor.test(shrubs_df$var_Grz07L06, shrubs_df$bite, method=c("pearson"))
## 
##  Pearson's product-moment correlation
## 
## data:  shrubs_df$var_Grz07L06 and shrubs_df$bite
## t = 2.5309, df = 41, p-value = 0.01531
## alternative hypothesis: true correlation is not equal to 0
## 95 percent confidence interval:
##  0.07558836 0.60152145
## sample estimates:
##       cor 
## 0.3675871
cor.test(shrubs_df$var_Grz07L06, shrubs_df$trump, method=c("pearson"))
## 
##  Pearson's product-moment correlation
## 
## data:  shrubs_df$var_Grz07L06 and shrubs_df$trump
## t = 0.3907, df = 41, p-value = 0.698
## alternative hypothesis: true correlation is not equal to 0
## 95 percent confidence interval:
##  -0.2439013  0.3547586
## sample estimates:
##        cor 
## 0.06090408

current grazing estimated through proxies is positively correlated with the difference in past to present grazing. It would even appear that in currently grazed areas grazing has even increased, and conversely, in today’s lightly grazed areas grazing has decreased.

grazing abandonment pattern and grazing facilities

Here we calculated the linear distance between our plots and points of interest for grazing (water points and grazing facilities). We only did this in an area where we were aware of all water points and grazing structures present. We calculated the distance linearly, although it would be more correct to account for slopes and natural obstacles to livestock movement. We tried using the “walktime” package (https://plugins.qgis.org/plugins/walkingtime/) but fortunately we could not obtain the distance expressed in “walktime” which takes into account obstacles, especially elevation. While the linear distance is therefore inaccurate, our results show a moderately significant correlation (p=0.065) between grazing abandonment (change in grazing between 2016 and ~1985) and distance to grazing points. Thus, the further distance the plot is from water points and grazing facilities, the more grazing abandonment occurred.

library(rgdal)
## Loading required package: sp
## Please note that rgdal will be retired by the end of 2023,
## plan transition to sf/stars/terra functions using GDAL and PROJ
## at your earliest convenience.
## 
## rgdal: version: 1.5-27, (SVN revision 1148)
## Geospatial Data Abstraction Library extensions to R successfully loaded
## Loaded GDAL runtime: GDAL 2.4.2, released 2019/06/28
## Path to GDAL shared files: /Library/Frameworks/R.framework/Versions/3.6/Resources/library/rgdal/gdal
## GDAL binary built with GEOS: FALSE 
## Loaded PROJ runtime: Rel. 5.2.0, September 15th, 2018, [PJ_VERSION: 520]
## Path to PROJ shared files: /Library/Frameworks/R.framework/Versions/3.6/Resources/library/rgdal/proj
## Linking to sp version:1.4-5
## Overwritten PROJ_LIB was /Library/Frameworks/R.framework/Versions/3.6/Resources/library/rgdal/proj
grazing_distance <- readOGR("grazing_linear_distance.gpkg")
## OGR data source with driver: GPKG 
## Source: "/Users/andreadetoma/mediterranean_alpine_shrubs/grazing_linear_distance.gpkg", layer: "grazing_linear_distance"
## with 260 features
## It has 80 fields
grazing_distance$Grz07L06 <-  grazing_distance$ndviL06 - grazing_distance$ndvi07

grazing_distance$var_Grz07L06 <- ((grazing_distance$Grz07L06) - (grazing_distance$past_Grz07E06))

cor.test(grazing_distance$var_Grz07L06, grazing_distance$Matrice.di.distanza_MEAN, method=c("pearson"))
## 
##  Pearson's product-moment correlation
## 
## data:  grazing_distance$var_Grz07L06 and grazing_distance$Matrice.di.distanza_MEAN
## t = 1.8604, df = 258, p-value = 0.06396
## alternative hypothesis: true correlation is not equal to 0
## 95 percent confidence interval:
##  -0.006691456  0.233442055
## sample estimates:
##       cor 
## 0.1150559