Visualization Playground

# Verify data periodicity
data <- DATA[iso3=="ken"]
data[, .N, by=.(sheet, period)]

##     sheet period     N
## 1: sheet1  month 10260
## 2: sheet1   year   855
## 3: sheet2  month 69480
## 4: sheet2   year  5790

Collection of visuals for WA Dashboard. Variable codes and definitions are in this shared catalog.

Water Availability

Basin Closure

(1 – (outflow¹ / gross inflow²)) * 100

# Yearly
dt <- data[id %in% c("outflow", "gross_inflow") & period=="year"]
dt <- dcast(dt, ...~id)[, value := 100 * (1 - (outflow / gross_inflow))]

plot_ts(dt, name="Pct. basin closure", unit="%", 
  title="Mara Basin Closure")

Availability per Capita

available water³ / population⁴

# Basin population from config
tot_pop <- ISO3[["ken"]]$population

# Yearly
dt <- data[id %in% c("available_water") & period=="year"]
dt <- dcast(dt, ...~id)[, value := 1E9 * (available_water) / tot_pop]

plot_ts(dt, name="Per Capita  Water Consumption", unit="m³",
  title="Mara Basin Per Capita Water Consumption")

Intra-annual summary below shows entire period statistics and 3 most recent years (can also be displayed as a polar/circular, see end of page).

# Monthly profile
dt <- data[id %in% c("available_water") & period=="month"]
dt <- dcast(dt, ...~id)[, value := 1E9 * available_water / tot_pop]

plot_profile(dt, unit="m³",
  title="Mara Basin Per Capita Water Consumption")

Water Available for Further Use

(utilizable outflow⁵)

# Yearly
dt <- data[id %in% c("utilized_flow") & period=="year"]
dt <- dcast(dt, ...~id)[, value := 1000 * utilized_flow]

plot_ts(dt, name="Utilizable Outflow", unit="MCM",
  title="Mara Basin Utilizable Outflow")

Water Use

Agricultural Water Use

((managed water use⁶ – non-ag ET⁷) / water consumed⁸)

# Yearly time-series
dt <- data[id %in% c("molu_et", "malu_et", "consumed_water",
  "molu_settlements_et", "molu_others_et", "malu_waterbodies_et", "malu_residential_et",
  "malu_industry_et", "malu_idomestic_et", "malu_iindustry_et", "malu_powerandenergy_et",
  "malu_others2_et")
  & period=="year"]

dt <- dcast(dt, year+month+date_end+date_start~id)[, value := 100 * (
  + molu_et + malu_et
  - molu_settlements_et - molu_others_et - malu_waterbodies_et - malu_residential_et
  - malu_industry_et - malu_idomestic_et - malu_iindustry_et - malu_powerandenergy_et
  - malu_others2_et
) / consumed_water]

plot_ts(dt, name="Ag. water use", unit="%",
  title="Agricultural water use")

# Monthly
dt <- data[id %in% c("molu_et", "malu_et", "consumed_water",
  "molu_settlements_et", "molu_others_et", "malu_waterbodies_et", "malu_residential_et",
  "malu_industry_et", "malu_idomestic_et", "malu_iindustry_et", "malu_powerandenergy_et",
  "malu_others2_et")
  & period=="month"]

dt <- dcast(dt, year+month+date_end+date_start~id)[, value := 100 * (
  + molu_et + malu_et
  - molu_settlements_et - molu_others_et - malu_waterbodies_et - malu_residential_et
  - malu_industry_et - malu_idomestic_et - malu_iindustry_et - malu_powerandenergy_et
  - malu_others2_et
) / consumed_water]

plot_profile(dt, name="Ag. water use", unit="%",
  title="Agricultural water use")

Environmental Stress

(percentage of time (months in the time series) during which environmental flows are not met) – Manohar/Mansoor to advise calculation from CSV WA output).

We use a low threshold value (e.g. < 0.05 km³). More details to follow.

threshold = 0.05

# Yearly time-series
dt <- data[id %in% c("reserved_outflow") & period=="month"]
dt <- dcast(dt, ...~id)[, value := reserved_outflow
][, .(
  date_end = max(date_end),
  date_start = min(date_end),
  value = 100 * sum(value <= threshold, na.rm=T)/.N
), by=.(iso3, year)]

plot_ts(dt, name="Insuffient Flow", unit="%",
  title="Mara Basin - Environmental Stress", 
  subtitle=sprintf("2003-2017 (%% of months in period below %s km³)", 
    threshold)
)

Basin Variability

Precipitation

(inter and intra-annual min and max over the time series)

Add option in dashboard for users to toggle units between volume and height?

# Monthly
dt <- data[id=="rainfall" & period=="month"]
dt <- dcast(dt, ...~id)[, value := rainfall]

plot_ts(dt, name="Precipitation", unit="km³", title="Precipitation")

plot_profile(dt, unit="km³", title="Precipitation")

Intra-annual variability can also be shown as a polar chart:

plot_profile(dt, unit="km³", polar=T, title="Precipitation")

Other Diagrams

Using a chord diagram to show water allocation between agricultural and non-ag. uses (or other breakdowns).

ag <- c("Irrigated crops", "Livestock and husbandry", "Greenhouses", 
  "Forest plantations", "Rainfed crops")

# (convoluted, include these constructed variables in ETL)
dt <- data[period=="year" & year==2017 &
    ((sheet=="sheet2" & id %like% "_et") | (sheet=="sheet1" & id %in% c("outflow")))
][META, on=.(id), `:=`(
  from = i.class,
  to = fcase(i.subclass %in% ag, "agriculture", default = "non-agriculture")
)][from != ""
][from=="OUTFLOW", to := "outflow"
][, from := "inflow"][, .(weight = sum(value, na.rm=T)), by=.(from, to)
][, `:=`(
  weight = weight/sum(weight, na.rm=T),
  color = pal[c("navy", "green", "orange")]
)]

plot_wheel(dt, rot=180, 
  colors=pal[c("blue", "navy", "green", "orange")],
  icons=c(`inflow`="tint", `agriculture`="envira", 
    `outflow`="water", `non-agriculture`="industry"),
  title="System Water Uses", subtitle="2017 (% allocation")

Using a Sankey graph to show ET allocation across land use categories (can be used to show a different combination of variables).

dt <- data[period=="year" & year==2017 & sheet=="sheet2" & id %like% "_et"
][META, on=.(id), `:=`(
  l1 = i.class,
  l2 = fcase(i.subclass %in% ag, "agriculture", default = "non-agriculture"),
  l3 = i.subclass
)][l1 != "", .(weight = sum(value, na.rm=T)), by=.(l1, l2, l3)
][, `:=`(
  weight = 100 * weight/sum(weight, na.rm=T)
)]

dt <- data_to_sankey(dt[, .(l1, l2, l3)])

plot_sankey(dt, 
  title="Evapotranspiration",
  subtitle="2017 (% allocation)")

Based on additional graphs found in WA+ basin reports.

# Yearly basin closure
dt <- data[id %in% c("outflow", "gross_inflow") & period=="year"]
dt <- dcast(dt, ...~id)

highchart(type="stock") %>%
  hc_add_series(dt, type="arearange", 
    hcaes(x=date_end, low=outflow, high=gross_inflow), 
    name="gross inflow", lineWidth=2, color=pal[[2]]
  ) %>%
  hc_add_series(dt, type="line", 
    hcaes(x=date_end, y=outflow), 
    name="outflow", lineWidth=2, color=pal[[1]]
  ) %>%  
  hc_add_series(dt, type="line", 
    hcaes(x=date_end, y=predict(lm(gross_inflow~year))), 
    name="trend", lineWidth=2, color=pal[[6]]
  ) %>%
  hc_add_series(dt, type="line", 
    hcaes(x=date_end, y=predict(lm(outflow~year))), 
    name="trend", lineWidth=2, color=pal[[6]], showInLegend=F
  ) %>%    
  th(
    title="Mara Basin Closure", 
    subtitle="Outflow / Gross Inflow (km³)")

# Monthly water balance (formula?)
dt <- data[id %in% 
    c("external_in", "landsc_et", "land_et", "q_gw_out", "q_sw_out", "delta_s") 
  & period=="month"]
dt <- dcast(dt, ...~id)[order(date_end), `:=`(
  inflow = external_in,
  outflow = landsc_et + land_et + q_gw_out + q_sw_out
  #delta_s = cumsum(delta_s)
)]

# Color zones
z <- dt[, .(
  x = date_end,
  zone = cumsum(inflow > outflow)
)][, .(x = last(x)), by=zone][, color := unname(pal[c(7, 2)])[1+odd(zone)]]
z <- lapply(1:nrow(z), function(i)  list(
  value = z[i, dt_tstp(x)], 
  color = z[i, color], 
  fillColor = alpha(z[i, color], .3)
))

highchart(type="stock") %>%
  hc_add_series(dt, type="arearange", 
    hcaes(x=date_end, low=outflow, high=inflow),
    name="Balance", lineWidth=2, zoneAxis="x", zones=z
  ) %>%
  # hc_add_series(dt, type="area",
  #   hcaes(x=date_end, y=delta_s),
  #   name="deltaS", lineWidth=2, color=pal[["red"]]
  # ) %>%
  th(
    title="Mara Basin Water Balance", 
    subtitle="Inflow / Outflow (km³)")

Verify definitions of inflow and outflow above.

BACOU, Melanie

Feb. 23, 2022