7_combined_protein_and_comparisons.Rmd

---
title: 'Comparisons: grazing livestock and crops'
author: "Johannes Piipponen"
date: "`r Sys.Date()`"
output: html_document
---

!! EDIT in revision
- "mixed" area analysis removed
- SI < 10 areas removed --> we now use SI < 33 areas to get more spatial coverage



Data saved to folder Intermediate inputs



# Calculate livestock contribution to total protein


Update: calculated only for current herd structure.

However, old codes below just in case one wants to explore this for different livestock species

Grazing livestock's share of total protein production on agricultural land
This map illustrates the percentage of protein produced by grazing livestock in relation to the combined protein production from both grazing livestock and the 27 SPAM food crops on agricultural land. It highlights the areas where grazing livestock plays a more significant role in contributing to the total protein production within agricultural systems.


# First, plot protein production levels
Both from AGB and SPAM, then also combined

```{r}

# calculate combined protein that we can get from a cell (combined = protein_grazing_livestock + protein_crops)
r_protein_combined_kg_perpix <-
  1000 * classify(r_prot_allcrops_sum_mt_perpix, cbind(NA, 0)) +
  classify(r_protein_from_AGB_kg_perpix$current_herd_str, cbind(NA, 0))

names(r_protein_combined_kg_perpix) <- "combined_protein"
#plot(r_protein_combined_kg_perpix, main = "combined protein")'
r_protein_combined_kg_perpix_0toNA <-
  classify(r_protein_combined_kg_perpix, cbind(0, NA))


pal_protein_lajolla <- scico(n = 6, palette = "lajolla", begin = 0.15, end = 1, direction = -1) 


# PER PIX
(plt_agb_and_crop_protein <-
  create_index_map(r_index = c( r_prot_allcrops_sum_mt_perpix,
                               r_protein_from_AGB_kg_perpix$current_herd_str/1000,
                               r_protein_combined_kg_perpix_0toNA/1000),
                   tocrs = "ESRI:54030",
                   index_main_title = "Protein production from AGB and from SPAM food crops in 2010",
                   index_label = "[kg/ha]",
                   colorpal = pal_protein_lajolla,
                   breakvals = c(0, 10, 50, 100, 250, 500, Inf),
                   breaknames = c("0-10","10-50", "50-100",
                                  "100-250","250-500",  ">500")))
# # interpretation:
# animals produce much less


```

# Grasing animals share of the total protein production

```{r}

# Grazing livestock's share of total protein production on agricultural land
# share of the total protein that originates from livestock
  ## first convert NA to 0
r_protein_from_AGB_current_str_kg_perpix_NAto0 <- 
  classify(r_protein_from_AGB_kg_perpix$current_herd_str, cbind(NA, 0))

r_share_of_total_protein_livestock <- 
  100* (r_protein_from_AGB_current_str_kg_perpix_NAto0 /
          r_protein_combined_kg_perpix)


qtm(r_share_of_total_protein_livestock, title = "Contribution of livestock to total protein production")
 # just shows that gl area >> cl area

pal_share <-  scico(n = 10, palette = "roma", begin = 0.1, end = 0.9, direction = -1) # change name?

create_index_map(r_index = r_share_of_total_protein_livestock , 
                     tocrs = "ESRI:54030",
                     index_main_title = "Grazing livestock share of tot protein prod",
                     index_label = "[%]",
                     colorpal = pal_share,
                     colorNA = "NA",
                     breakvals = c(0, 10, 25, 50, 75, 90, 100),
                     breaknames = c("<10", "10-25", "25-50","50-75","75-90", "90-100"))




# ------------------------------------------------------ mask these to key areas (defined in another script called)
  # over 90% grazing lands
# r_share_of_total_protein_livestock_whereover90gl <- # not sure if this is relevant
#   mask(r_share_of_total_protein_livestock,
#        r_gl_share_of_agrland_over90)

  # over 90% croplands
# r_share_of_total_protein_livestock_whereover90cl <- # not sure if needed
#   mask(r_share_of_total_protein_livestock,
#        r_cl_share_of_agrland_over90)


  # 10% of over (of agr land) croplands AND 10% of over (of agr land) grazing lands
# r_share_of_total_protein_livestock_mixed <-
#   mask(r_share_of_total_protein_livestock,
#        r_gl_cl_share_mixed)






# --------------------------------------------------- same masked to cl
  ## cropland mask 
  ## share of livestock masked to croplands to show that on cropland areas livestock produce only a little
# r_share_of_total_protein_livestock_croplandmask <-
#   mask(r_share_of_total_protein_livestock, 
#        r_fraction_cl_0toNA)
# 
# plot(r_share_of_total_protein_livestock_croplandmask, 
#      main = "Contribution of livestock to total protein production on croplands")



# ----------------------- further masked to cl where SI extremely low (below 10)

  ## SI below 10 mask -- must also be larger than 0
# SI_5arcmin_range_1to10 <- SI_5arcmin # check what "source" row says -- should be rcp rainfed
# SI_5arcmin_range_1to10[SI_5arcmin_range_1to10 >= 10] <- NA
# SI_5arcmin_range_1to10[SI_5arcmin_range_1to10 <= 1] <- NA

## 1<=SI<33 
# SI_5arcmin_range_1to33 <- ifel(SI_5arcmin_range_1to33 >= 1 & SI_5arcmin_range_1to33 < 33, 1, NA)
# 
# plot(SI_5arcmin_range_1to33, main = "1<= SI < 33")
# 
# 
# r_share_of_total_protein_livestock_SIbetween1to33mask <-
#   mask(r_share_of_total_protein_livestock, 
#        SI_5arcmin_range_1to33)
# 
# plot(r_share_of_total_protein_livestock_SIbetween1to33mask)

# SI_5arcmin_below10_croplandmask <- mask(SI_5arcmin_below10, r_fraction_cl_0toNA) # no need to show this - use all agr land instead?
# plot(SI_5arcmin_below10_croplandmask, main = "SI below 10 on croplands")
# 
# r_share_of_total_protein_livestock_SIbelow10mask_croplandmask <-
#   mask(r_share_of_total_protein_livestock, SI_5arcmin_below10_croplandmask)
# 
# plot(r_share_of_total_protein_livestock_SIbelow10mask_croplandmask, 
#      main = "Contribution of livestock to total protein production on croplands where SI < 10")
```




# Protein kg per ha for figures

```{r}
# total combined protein per ha
r_protein_from_AGB_current_str_kg_ha_NAto0 <-
  classify(r_protein_from_AGB_kg_ha$current_herd_str, cbind(NA,0))
  
r_prot_allcrops_sum_kg_ha_NAto0 <-
  classify(r_prot_allcrops_sum_kg_ha, cbind(NA,0))


r_protein_combined_kg_ha <- 
  r_protein_from_AGB_current_str_kg_ha_NAto0 +
  r_prot_allcrops_sum_kg_ha_NAto0

r_protein_combined_kg_ha <-
  mask(r_protein_combined_kg_ha, 
       r_fraction_gl_cl_total_0toNA) # focus only on our study area

# total combined protein per ha - masked to key cropland areas
# r_protein_combined_kg_ha_whereover90cl <-
#   mask(r_protein_combined_kg_ha, 
#        r_cl_share_of_agrland_over90)

# total combined protein per ha - masked to mixed areas
# r_protein_combined_kg_ha_wheremixed <-
#   mask(r_protein_combined_kg_ha, 
#        r_gl_cl_share_mixed)

# total combined protein per ha - croplandmask
# r_protein_combined_kg_ha_croplandmask <-
#   mask(r_protein_combined_kg_ha, 
#        r_fraction_cl_0toNA)

# total combined protein per ha - outside croplandmask -- same as only livestock here
# r_protein_combined_kg_ha_outside_croplands <-
#   mask(r_protein_from_AGB_kg_ha$current_herd_str, # could be also r_protein_combined_kg_ha but this is more correct
#        r_fraction_cl_0toNA, inverse = T)

# further masked to cl where SI is marginal
r_protein_combined_kg_ha_SIbetween1to33mask <-
  mask(r_protein_combined_kg_ha,
       SI_5arcmin_range_1to33)

plot(r_protein_combined_kg_ha_SIbetween1to33mask)






```



# Global sums 

```{r}
# Total combined protein
global(r_protein_combined_kg_perpix, "sum", na.rm = T)/1e9 # 385 in 2020, was 285 mmt in 2010

# Total combined protein on key cropland areas
# r_protein_combined_kg_perpix_whereover90cl <-
#   mask(r_protein_combined_kg_perpix, 
#        r_cl_share_of_agrland_over90)
# global(r_protein_combined_kg_perpix_whereover90cl, "sum", na.rm = T)/1e9 # 107


# Total combined protein on mixed areas
# r_protein_combined_kg_perpix_wheremixed <-
#   mask(r_protein_combined_kg_perpix, 
#        r_gl_cl_share_mixed)
# global(r_protein_combined_kg_perpix_wheremixed, "sum", na.rm = T)/1e9 # was 170

# Total combined protein on croplands
# r_protein_combined_kg_perpix_croplandmask <-
#   mask(r_protein_combined_kg_perpix, 
#        r_fraction_cl_0toNA)
# global(r_protein_combined_kg_perpix_croplandmask, "sum", na.rm = T)/1e9 # 281

# Total combined protein on croplands where 1<= SI < 33
r_protein_combined_kg_perpix_SIbetween1to33mask <-
  mask(r_protein_combined_kg_perpix, 
       SI_5arcmin_range_1to33)
global(r_protein_combined_kg_perpix_SIbetween1to33mask, "sum", na.rm = T)/1e9 #55.6

# Total combined protein outside croplands -- this should be same as amount of protein current herd produces outside cl
r_protein_combined_kg_perpix_outside_croplands <-
    mask(r_protein_combined_kg_perpix,
         r_fraction_cl_0toNA, inverse = T)
global(r_protein_combined_kg_perpix_outside_croplands, "sum", na.rm = T)/1e9 # 6.7 -- 4

mask(r_protein_from_AGB_kg_perpix$current_herd_str,
     r_fraction_cl_0toNA, inverse = T) %>% 
  global(., "sum", na.rm = T)/1e9 # 6.7 -- was 4





```

# Global sums in broader context

```{r}
  # combined protein from crops and livestock = 385.1329 mmt. This is 
(1e12 * global(r_protein_combined_kg_perpix, "sum", na.rm = T)/1e9) /
   6869581990 / # persons
  365 
# 154 --> was114 g/capita/day


# how many people can be fed with this?
#((1e12 * global(r_protein_combined_kg_perpix, "sum", na.rm = T)/1e9) / # grams (mmt = 1e9 kg)

# ((1000 * global(r_protein_combined_kg_perpix, "sum", na.rm = T)) / # grams of protein produced
#   (1000* foodBalanceSheets_semiraw %>% filter(Area == "World", Item == "Population") %>% pull(Y2020)) / # number of persons in 2020
#   365 ) /# days in a year
#   (0.8*70) # protein need in 
# 

```




# Save rasters

```{r}
# -------------------------------------------- save rasters we need
writeRaster(r_protein_combined_kg_ha,
            filename = here("Data", "Intermediate_input", "r_protein_combined_kg_ha.tif"),
            overwrite = TRUE)

# writeRaster(r_protein_combined_kg_ha_whereover90cl,
#             filename = here("Data", "Intermediate_input", "r_protein_combined_kg_ha_whereover90cl.tif"),
#             overwrite = TRUE)

# writeRaster(r_protein_combined_kg_ha_wheremixed,
#             filename = here("Data", "Intermediate_input", "r_protein_combined_kg_ha_wheremixed.tif"),
#             overwrite = TRUE)
 
writeRaster(r_protein_combined_kg_perpix,
            filename = here("Data", "Intermediate_input", "r_protein_combined_kg_perpix.tif"),
            overwrite = TRUE)


# writeRaster(r_protein_combined_kg_ha_croplandmask,
#             filename = here("Data", "Intermediate_input", "r_protein_combined_kg_ha_croplandmask.tif"),
#             overwrite = TRUE)
writeRaster(r_protein_combined_kg_ha_SIbetween1to33mask,
            filename = here("Data", "Intermediate_input", "r_protein_combined_kg_ha_SIbetween1to33mask.tif"),
            overwrite = TRUE)


################################## shares
writeRaster(r_share_of_total_protein_livestock,
            filename = here("Data", "Intermediate_input", "r_share_of_total_protein_livestock.tif"),
            overwrite = TRUE)
# writeRaster(r_share_of_total_protein_livestock_croplandmask,
#             filename = here("Data", "Intermediate_input", "r_share_of_total_protein_livestock_croplandmask.tif"),
#             overwrite = TRUE)
# writeRaster(r_share_of_total_protein_livestock_mixed,
#             filename = here("Data", "Intermediate_input", "r_share_of_total_protein_livestock_mixed.tif"),
#             overwrite = TRUE)


writeRaster(r_share_of_total_protein_livestock_SIbetween1to33mask,
            filename = here("Data", "Intermediate_input", "r_share_of_total_protein_livestock_SIbetween1to33mask.tif"),
            overwrite = TRUE)


# writeRaster(SI_5arcmin_range_1to10,
#             filename = here("Data", "Intermediate_input", "SI_5arcmin_range_1to10.tif"),
#             overwrite = TRUE)
```