The present precipitation over the Bolivian and Peruvian altiplano is modulated by topography and variations in the reflectivity of the surface. Together with the diurnal variation of solar radiation these produce atmospheric circulations ("diurnally-varying circulations") that in term produce preferred regions of ascent and descent. These regions of ascent and descent modulate the evolution of clouds and precipitation. In the recent geological past (< 15,000 years) the large paleolake Tauca existed on the southern altiplano. The effect of the lake on the precipitation climatology has not been considered in detail in the climate literature.

Figure 1. The left panel is a Modis image of the Altiplano showing Lake Titicaca (1), Lake Poopo (2), Salar de Coipasa (3) and Salar de Uyuni (4). This is how the Altiplano looks at present times from the air. The right panel shows a map of the Altiplano indicating the extension of Paleolake Tauca in light blue. The current lakes and dry salt flats are also inficated with contours and shaded with lines. The edges of the basin are indicated with a continuous black line.

We hypothesize that the existence of the lake profoundly changed the local precipitation climatology of the southern altiplano. This we believe occurred primarily through the generation of vertical motion on the mesoscale that produced regular nocturnal convection over the lake and affected the mean balance between precipitation and evaporation. We further hypothesize that the change from the currently dry conditions over the Salar de Uyuni to a salar covered with only several m of water would have significantly affected the hydrologic balance of the Salar de Uyuni and would have resulted in a growing lake depth, independent of small large-scale climate fluctuations. The may be an unstable equilibrium such that if a sufficiently large number of wet years occur in succession, the Salar will continue to grow by self-induced mesoscale precipitation.

Although all of these hypotheses can only be investigated through model simulations, we carried out series of measurements to help constrain the model simulations and to serve as verification data for the validity of the mesoscale numerical weather prediction model. The results from the observations are intended to confirm that the mean vertical motions over Lake Titicaca are upward and are downward over the highly reflective Salar de Uyuni. The hypothesized much stronger early morning upward motion over the Lake compared with the Salar were measured/calculated with a polygon of frequent pilot balloon observations and integration of the continuity equation. Variations of stability throughout the diurnal cycle were measured by tethered balloon-based radiosonde observations. Details of the rainfall "climatology" were measured by a dense network of daily-read rain gauges that maintained during the entire wet season.

Lake Titicaca and Salar de Uyuni suggested that they were altering the climate of their environments in many ways. This was revealed through satellite imagery, and climatologies of temperature and rainfall. The Salar de Uyuni, given its high reflectivity and relatively low heat capacity when compared to water, results in less rainfall over the salar as a response of downwelling salar-induced motions.

Lake Titicaca, in contrast, absorbs larger amounts of solar radiation since the surface albedo is lower, and keeps the water temperature almost constant during the day given the depth of the lake (~ 300 m in the deepest part) and the large heat capacity of water. This results in an important horizontal temperature contrast over the lake shore regions during the night which leads to the development of land-lake nocturnal breezes, low-level convergence and mesoscale convective storms over the lake. These storms donot develop every night and they seem to depend on the vertical moisture profile over the Altiplano and synoptic conditions within others. Even though these storms have relatively small sizes and do not develop very much in the vertical (i.e. they are not very evident in the satellite imagery), they produce large amounts of rainfall which result in almost twice as much annual rainfall over and in the shores of Lake Titicaca than in the surrounding terrain.

Figure 2. The left panel shows the frequency of temperatures colder than -13°C between 8pm and 10am obtained from GOES-8 IR4 Channel satellite images collected during the SALLJEX. The right panel is a scheme showing the development of mesoscale convection over Lake Titicaca aligned with the axis of convergence of the breezes.

Figure 2 shows an map of the Altiplano indicating in shaded the frequency of temperatures colder than -13°C between 8pm and 10am obtained from GOES-8 IR4 Channel satellite images collected during the SALLJEX (Dec.1.2002 - Feb.28.2003). This shows the development of mesoscale convective storms over the lake and over the eastern slopes of the Andes during this period. The image also indicates that over Salar de Uyuni the nocturnal convection is not frequent or may be even absent. Another indication of larger rainfall rates over the lake than over land is indicated by figure 3. This is the difference between the monthly rainfall climatologies of Puno and Juliaca, both in the Peruvian side of the Altiplano and located at only 40 km one from the other. Puno is located by the lake whereas Juliaca is located about 35km inland from Titicaca. Puno receives about 100 mm of rainfall more than Juliaca, and this excess is recorded mainly between December and April, which corresponds to the rainy season. This suggests that lake-effects storms mainly or only occur during this period, particularly in February and March.

Figure 3. Difference in the Monthly Rainfall observed in Puno and Juliaca, both located in the Peruvian side of the Altiplano. Puno lies at the shores of Lake Titicaca and Juliaca about 35 km inland.

Why setting an enhanced raingauge network?

Even though previous authors have sketched rainfall analyses showing a maximum over the lake, the number of stations used for the analyses has been very reduced to reproduce a fair analysis. Considering that the nature of the precipitation in the Altiplano is convective which leads to large variations in the horizontal, a dense raingauge network of at least 10 km between the stations was needed to describe the field with reasonable accuracy. With rainfall analyses in general, the densest the network the best the analysis.

Figure 4. Rainfall Analyses in the vicinity of Lake Titicaca. The left panel shows the mean annual rainfall in mm accumulated from 1957 to 1961 extracted from Kessler and Montheim (1968) after Scherdtfeger (1976) and reformatted. The right panel shows a 3-month rainfall analysis using SALLJEX data. The white lines show the regions near the lake where the density of the networks and therefore quality of the analyses differ the most. NOTE: The scale of the contours differ.

Figure 4 shows the difference between two rainfall analyses in the Lake Titicaca region. The left panel corresponds to an annual rainfall analysis by Kessler and Montheim (1968) after Scherdtfeger (1976) and reformatted. The right panel corresponds to the December-February rainfall analysis generated with the SALLJEX raingauge network. Even though the periods are different, note the difference in the resolution of the analyses near the lake. This helps to understand why taking advantage of the large number of observations that were programmed for the SALLJEX field campaign the initially designed raingauge network was extended to the Altiplano region with higher densities near and over Lake Titicaca.

Last Update: Feb 26, 2005
Comments to jose.galvez@noaa.gov.

Last Update: March 2, 2005