Supplementary Material 1: Water Routing in the Lake Catchment

Supplementary Material 1: Water Routing in the Lake Catchment

Electronic Supplementary Material

Supplementary Material 1: water routing in the lake catchment

Water routing in the catchment of Lake Chichój is quite complex due to flow redistributions in its carbonate basement. Based on the surface topography, the lake possesses a drainage area of 21.0-23.5 km2 (Fig.4). The lake is fed by the Chijuljá River, which provides more than 95% of the surface water input (Alpizurez-Palma 1978) but drains only 50% of the topographic catchment. Two springs or more bring water, directly to the lake or into the surrounding wetlands (Fig.1). A tributary at the eastern end of the lake drains a patch of wetlands located further east, but these eastern wetlands are also drained by a stream that flows to the north and shortcuts the lake (Fig. 4). This configuration allows for the eastern tributary to flow backwards into the wetlands during lake highstands (Alpizurez-Palma 1978). The rest of the catchment drains into the lake either diffusively or through unknown lacustrine springs, unless water merely bypasses the lake. In spite of this complexity, the Chijuljá River seems to be the main provider with a discharge of 0.53 m3/s in November 1977 (Alpizurez-Palma 1978). This is in agreement with a mean annual discharge of 0.54 m3/s calculated using a catchment-averaged mean annual rainfall of 1500 mm/y (data from INSIVUMEH, regional precipitation map), assuming full conversion to runoff.

Upon leaving the lake, water routing is equally complex. A single discharge estimate at the lake outlet (Alpizurez-Palma 1978) yielded a significantly lower discharge (0.33m3/s), suggesting subaquatic leakage. If leakage does occur, it could feed large sulfate-bearing springs located along the Polochic fault, two kilometers south of the lake (Fig.3). The discharge of these springs is much larger than the amount of precipitation their upslope catchments can provide, and among the conterminous catchment of these springs, Lake Chichój stands 220-630 m higher than the springs. There is therefore a strong potential for lake water to migrate to these springs by karstic redistribution through a ridge made of Permian limestones. The high sulfate content of the springs indicates that their water has interacted with gypsum, also pointing to a connection with the Chichój catchment (see Supplementary Material 2).

Supplementary Material 2: the role of gypsum in the formation of Lake Chichój

The lake bathymetry shows that the lake results from the impoundment of water into an alignment of at least three karstic depressions. Normal faults inrecent stream sediments at the eastern end of the lake indicate that subsidence is recent or active (Fig.1).

Considering that other karstic depressions in the highlands do not hold permanent water bodies, gypsum dissolution at depth may cause the existence of the lake. Not far from Lake Chichój is seen the most extensive outcrop of gypsum in Central America. This Late Jurassic-Early Cretaceous gypsum (Donnelly et al. 1990) daylights at the ground surface over 40 km (Fig.3). The dolines occupied by the lake are aligned along strike with the gypsum body, and are situated 5 km to the west of its easternmost surface appearance. South of the lake, a series of large sulfate-rich springs along the trace of the Polochic fault reveals the presence of gypsum at depth in the vicinity of Lake Chichój (Fig.3). The geological setting strongly argues for the presence of gypsum under the lake.

Considerations about the rate of ground subsidence also support a genesis by gypsum dissolution: the lake is inset into pumice deposits of the Los Chocoyos formation, an ash-flow deposit that was emplaced 84 ± 5 ky ago (Drexler et al. 1980; Rose et al. 1986). Foundering of the pumice fill into the dolines can be estimated to have occurred at a rate of at least 0.55 mm/y, the rate necessary to bring the top of the pumice layer, 15 m above the lake, down to the bottom of the deepest doline (-32 m). This is a minimal estimate, considering that lake sediments are also filling the dolines. Such a subsidence is much higher than the average 0.03 mm/y dissolution rate calculated for the surrounding carbonates based on solute exports in rivers (Corbel 1959). The subsidence rate is also higher than any known limestone dissolution rate in subtropical wet mountainous areas receiving similar amounts of rainfall (1500 mm/y). On the other hand it is compatible with reported dissolution rates for gypsum (Ford and Williams 2007). Such a high rate may maintain the accommodation space necessary for the lake to persist over time, since measured sedimentation rates are very high. Upon infilling, the lake would then resemble the surrounding wetlands.

The flatness of the wetlands contrasts with the steep topography of the surrounding karst towers and with the steep dolines of the lake. The wetlands are graded to the lake level and thus ultimately to the elevation at which the lake outlet exits the catchment into the carbonates. The wetlands are likely former collapse structures, similar to the modern lake, filled by sediments up to the level at which water exists the lake region. We observed laminated sediments at the base of a 1.3 m-long core extracted from the wetlands (core 10P14, Fig.1): they attest the presence of open water conditions in the past in the wetlands. Lake Chichój is either most recent or most long-lived manifestation of ground subsidence. Dry valleys at the north-east end of the lake catchment (Fig.4) are the beheaded remnants of a preexisting array of north-flowing streams that were obliterated now-filled collapsed structures of similar origin.

The susceptibility of the lake and of its surroundings to ground collapse represent a significant hazard for the town of San Cristóbal, which could be affected by soil collapse and soil failures toward the lake.

References

Corbel (1959) Erosion en terrain calcaire (vitesse d’érosion et morphologie). Annales de Géographie 366: 7-120

Donnelly TW, Horne GS, Finch R, López-Ramos E (1990) Northern Central America: The Maya and Chortís blocks. In:Dengo G, Case JE (eds) The Geology of North America, H: The Caribbean Region. Boulder, CO, Geol Soc Am: 37-76

Ford D, Williams P (2007) Karst Hydrogeology and Geomorphology. John Wiley & Sons, Chichester, UK