Abstract
In the Colombian Caribbean region, there are few studies that evaluated the behavior of one of the most commonly used variables in hydrological analyses: the maximum daily rainfall (Pmax-24h). In this study, multiannual Pmax-24h time series from 19 rain gauges, located within the department of Atlántico, were analyzed to (a) determine possible increasing/decreasing trends over time, (b) identify regions with homogeneous behavior of Pmax-24h, (c) assess whether the time series are better suited under either a stationary or non-stationary frequency analysis, (d) generate isohyetal maps under stationary, non-stationary, and mixed conditions, and (e) evaluate the isohyetal maps by means of the calculation of areal rainfall (Pareal) in nine watersheds. In spite of the presence of both increasing and decreasing trends, only the Puerto Giraldo rain gauge showed a significant decreasing trend. Also, three regions (east, central, and west) with similar Pmax-24h behavior were identified. According to the Akaike information criterion test, 79% of the rain gauges showed better fit under stationary conditions. Finally, statistical analysis revealed that, under stationary conditions, the errors in the calculation of Pareal were more frequent, while the magnitude of the errors was larger under non-stationary conditions, especially in the central–south region.
Highlights
Changes in the pattern of behavior of hydrometeorological variables are influenced, among others, by population growth, watershed land-use/land-cover (LULC) changes, and the increase in greenhouse gases emissions [1,2].In Colombia, the Institute of Hydrology, Meteorology, and Environmental Studies (IDEAM, in Spanish) conducted several studies focused on evaluating the changes in the behavioral patterns of some hydrometeorological variables [3,4,5,6,7]
When the stationary and non-stationary conditions were compared one-to-one, it was observed that, in 57.8% of the cases (26 out of 45), the REr values for the non-stationary conditions were lower than their stationary counterparts. These results suggest that (a) the error might be more frequent when using the stationary condition isohyetal maps, and (b) additional attention should be paid during the design of hydraulic structures under stationary frequency analysis, especially as it was found that this scenario tends to underestimate the Pareal (Table 10)
The regionalization will be of great help for the Pmax-24h analysis in ungauged areas given the fact that the department of Atlántico is, among the remaining six departments of the Colombian Caribbean region, the one with the lowest rain gauge density
Summary
IDEAM [3] analyzed the annual average rainfall trend over the periods 2011–2040, 2041–2070, and 2071–2100 for the different departments (political and administrative territorial units) that compose the five regions of Colombia (Caribbean, Pacific, Andean, Orinoco, and Amazon). It was predicted that some of the municipalities located in the southeast of the department will be the most affected. Such changes in the hydrological cycle can lead to (a) decreases in the water supply (both for human consumption and for the different sectors of the economy), (b) possible water supply cost increase, Water 2019, 11, 2453; doi:10.3390/w11122453 www.mdpi.com/journal/water
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
