For my PhD, I studied the palaeolimnology of two lakes in the Ethiopian rift valley, using diatoms to reconstruct changes in the water chemistry of Lake Awassa, an oligosaline caldera lake which retains its low salinity despite having no effluent rivers, and Lake Tilo, one of three alkaline maar lakes close to the Bilate River toward the western edge of the rift valley.
At the time, crater lakes were widely thought of as giant rain gauges, but, of course, groundwater plays a large role in controlling lake depth and salinity, complicating the relationship these variables and climate. Given the importance of groundwater, I wanted to know the elevation of Lake Tilo, the neighbouring lakes Budemeda and Mechefera, and the Bilate River to better understand how groundwater would flow between them. Unfortunately, the 10m contour resolution on the available map was insufficiently precise and I was hopeless at using stereo-photographs to estimate elevations, so this never got done. But now I don’t need to use stereo-photographs, I can download a high-resolution digital elevation model — SRTM data — from https://earthexplorer.usgs.gov/ and plot it.
Here, I am using the 1′ resolution data. Each tile is about 25MB and covers a 1°x1° area.
library(tidyverse)
tilo = raster::raster(x = "n07_e038_1arc_v3.tif")
lakes = data_frame(
lake = factor(c("B", "T", "M"), levels = c("B", "T", "M")),
x = c(38.09417, 38.09417, 38.08500),
y = c(7.096667, 7.065000, 7.042500)
)
tilo2 = raster::as.data.frame(tilo, xy = TRUE) %>%
rename(elevation = 3)
tilo2 %>%
filter(between(x, 38.04, 38.15), between(y, 7.01, 7.12)) %>%
ggplot(aes(x = x, y = y, fill = elevation)) +
geom_raster() +
scale_fill_continuous(type = "viridis") +
scale_x_continuous(expand = c(0, 0)) +
scale_y_continuous(expand = c(0, 0)) +
coord_equal() +
geom_text(aes(x, y, label = lake), lakes, colour = "white", inherit.aes = FALSE) +
labs(fill = "Elevation, m", x = "°E", y = "°N")
Budemeda is the northernmost lake in a deep, steep-sided crater. Tilo is the middle lake. Mechefera is the southernmost lake in a shallow crater. When I visited, Mechefera supported a large population of flamingoes. Wading past the dead and dying flamingoes on the shores of the small cider-cone island was not pleasant. They kept the resident hyaena fed though.
Clik here to view.
DEM of the three lakes.
To my surprise, there is a smaller fourth crater south-west of Lake Mechefera – I thought this was another cinder cone in the east-west line of cones. Google maps shows the crater floor to be vegetated – I suspect it is swampy and may be seasonally flooded.
Now I can plot some E-W profiles through the lakes.
`%nearest%` <- function(a, b){#find nearest value#slow
a1 <- unique(a)
res <- sapply(b, function(i){ a == a1[which.min(abs(a1 - i))] }) apply(res, 1, any) } tilo2 %>%
filter(between(x, 38.04, 38.15)) %>%
filter(y %nearest% lakes$y) %>%
mutate(lake = factor(y, labels = c("M", "T", "B"))) %>%
ggplot(aes(x = x, y = elevation, colour = lake, label = lake)) +
geom_path() +
scale_colour_discrete(limits = c("B", "T", "M")) +
directlabels::geom_dl(method = "first.points") +
directlabels::geom_dl(method = "last.points") +
labs(y = "Elevation, m", colour = "Lake", x = "°E") +
theme(legend.position = "none")
Clik here to view.
East-west profiles through the three lakes.
Lake Budemeda appears to be 13 m below the elevation of the river immediately to the west. Lake Tilo is a couple of metres lower than Lake Budemeda, but less that five metres below the river to the west. Lake Mechefera is about 15 m below Lake Tilo and 8 m below the river immediately to the south. This means that river water will tend to percolate from the river into all three lakes, but especially Budemeda and Mechefera, and that water will flow from Budemeda to the lakes to the south. This probably partially explains the high salinity of the thermal springs along the northern shore of Lake Tilo (I should have realised that when I was doing my thesis).
Lacustrine sediment high on the crater walls show that during the early Holocene the water level in Tilo was about 30m higher that at present, far above the river level. At this time, the lake was fresh and there was rapid accumulation of diatomite dominated by Aulacoseira granulata.
SRTM data are a great resource, especially for closed basin lakes.