Regional Regression Models of Mean Annual Streamflow and Design Flow in a Tropical Region of Colombia.

Abstract

A responsible and sustainable water resources management is important. It is also crucial to take the presence of watercourses into account during design of hydraulic structures such as small dams and bridges. Doing this requires knowledge of estimates of annual streamflow and extreme flows to be used for design purposes. This study is divided into two parts, both focusing on the departments of Santander and Norte de Santander in Colombia where a tropical climate dominates. Both studies treat the creation of regression models using statistical tests to secure their reliability. The first study creates models for estimation of mean annual streamflow by relating streamflow to geomorphic and climate characteristics. Models are created for each individual department as well as for smaller, hydrologically homogeneous regions. The majority of the results show remarkably good fits with adjusted R2-values ranging from 86.1-99.3 % and VIF-values in the range of
1.00-3.20. The second study creates regression models for estimation of design flows, with return periods of 10, 20, 50 and 100 years, using geomorphic variables. It is found that watershed area, elevation and slope result in the best fit giving adjusted R2-values in the range of 72.8-75.7 % and a VIF-value equal to 1.536.

Popular Abstract

Simple Models to Predict Streamflow in Tropical Colombia
As climate changes and clean water becomes scarce the need for sustainable water resources
management is of high importance. Hydrological models to predict streamflow are therefore a
tool which is important to develop for all parts of the world.
By gathering simple variables about climate and geography for a water basin, linear models can be
created to explain the patterns of streamflow in the area. This has been done for decades in
non-tropical areas while information for tropical regions has been more scarce. Finally it has been
proven that models can be created for tropical regions as well. In fact only two variables are needed
to obtain a good model; area and precipitation. Such models were found to be useful to predict
extreme flows as well. In this case, extreme flow values can be explained using other variables, such
as information about water basin slope and elevation proved to be useful.
The global climate changes and extreme weather events become more common, leading to
problems in water management, both in forms of extreme drought and extreme flooding. These
droughts can lead to water scarcity, with catastrophic consequences for the population. At the same
time, unexpected extreme flows can have catastrophic consequences in forms of dam breaks or
bridge failures. The need for knowledge of water access and flows is therefore of greater importance
than ever.
The models created can be used in the water resources management to create a sustainable use of
water. It is important to know how much water the industries can use without depriving the water
that is used for drinking and sanitation. Further the well-being of the environment and the
ecosystems are of high importance as well, therefore we have to make sure that enough water is left
in the streams for the nature to thrive.
The resulting models will also be an important tool for Colombian decision makers when making
decisions regarding the water use for these tropical regions . As mentioned, this will contribute to a
more sustainable management of Colombian water resources and will minimize the risks of future
water scarcity. Statistically speaking, water scarcity really shouldn’t be a problem for Colombia as it
is one of the countries with the highest amount of water per inhabitant. However, due to the
distribution of the water in the country, bad infrastructure and pollution this is an impending threat.
The models are also created for extreme flows and as the country is in need of better infrastructure
they will serve as an important tool for the construction of infrastructure such as bridges and smaller
dams and ponds.
The use of models will also mean a more efficient and reliable way of working and obtaining
information about flow. In the older days this information was collected manually from stations in
the stream, where the changes in flow had been registered. This was of course very inefficient and
information could often be lost. In today's society the measurement stations are more efficient,
although they are also decreasing. Therefore models like these ones, that requires only a few simple
measured variables and that can function well without a long historical data are needed. They are
easy to work with at the same time as they are efficient and reliable.

• author: Zigaf Elfellah, Anisa and Eriksson, Amanda
• supervisor: Magnus Persson; Edgar Leonardo Villarreal González
• organization: Division of Water Resources Engineering
• course: VVRM05 20172
• year: 2018
• type: H2 - Master's Degree (Two Years)
• subject: Technology and Engineering
• keywords: mean annual streamflow, extreme flow, design flow, regional regression model, ungauged watershed, hydrological homogeneous region, Gumbel distribution
• publication/series: TVVR18/5003
• report number: 18/5003
• ISSN: 1101-9824
• language: English
• additional info: Both authors contributed equally to the article.; Examiner:Linus Zhang
• id: 8936276
• date added to LUP: 2018-02-21 15:04:23
• date last changed: 2018-02-21 15:04:23