Hydrograph separation using hydrochemical tracers in the Makanya catchment, Tanzania

doi:10.1016/j.pce.2007.04.015

Physics and Chemistry of the Earth, Parts A/B/C

Volume 33, Issues 1–2, 2008, Pages 151-156

https://doi.org/10.1016/j.pce.2007.04.015 Get rights and content

Abstract

Hydrochemical tracers were used to separate and quantify different runoff components in the semi-arid Makanya catchment in the South Pare Mountains of Tanzania. One flood event was investigated during the rainy season of October–December 2005 and analysed for electrical conductivity, dissolved silica and major anions and cations. The event on 9 November 2005 showed two peaks, each originating from one of two sub-catchments, upper-Vudee and Ndolwa, each with a distinct water quality signature. Hydrograph separation indicated that the two peaks in the hydrograph originated from a delay in response between the two catchments. The hydrograph separation indicated that, for this event, over 95% of the discharge could be attributed to sub-surface runoff, while the remainder was due to faster surface runoff processes. The dominance of sub-surface processes was also indicated by the lack of suspended sediments in the samples, which is a clear indication that no surface runoff took place.

Introduction

In many parts of the world catchments are not or poorly gauged. In particular in developing countries, catchments are predominantly ungauged, as a result of lack of adequate resources (Mazvimavi, 2003). There are two main factors in sub-Saharan Africa that affect the predictability of hydrological responses. First, the climatic variability is both spatially and temporally very high. Additionally, increasing population densities and the resulting dynamic land use changes affect hydrological responses.

Chemical hydrograph separation is a method to define the origin and composition of the runoff during floods (Uhlenbrook et al., 2002). This method is based on the mixing of two or more water types with known and distinct hydrochemical characteristics, where the ratio of mixing determines the concentrations in the stream. Chemical hydrograph separation is predominantly done to separate sub-surface and surface runoff. Pre-event (“old”) and event (“new”) water can be separated using environmental isotopes (Hooper and Shoemaker, 1986). Uhlenbrook et al. (2002) also used a chemical method for a separation between event water, shallow groundwater, and deep groundwater, using isotopes and dissolved silica (3-component separation). Hydrochemical tracers, such as Ca⁺, Mg²⁺, ${SO}_{4}^{2 -}$ and Cl⁻ are relatively inexpensive to analyse and easy to use for hydrograph separation (Ribolzi et al., 2000, Soulsby et al., 2004, Tardy et al., 2004, Wels et al., 1991). The main challenge with using these tracers is that the original concentrations can change along the specific flow path (non-conservative behaviour). With a limited amount of time between rainfall and runoff, this can be neglected. Although more expensive, hydrograph separation can also be done using isotopes, or a combination of isotope and chemical tracers (Sklash and Farvolden, 1979, Buttle, 1994, Ladouche et al., 2001, Marc et al., 2001).

Chemical hydrograph separation has been mostly applied in humid temperate climates. This paper describes the application of this method using hydrochemical tracers in a semi-arid meso-scale catchment. The Makanya catchment is located in the South Pare Mountains, northern Tanzania. It is part of the Pangani river basin, one of the nine river basins in Tanzania (Fig. 1, Fig. 2). It is a poorly gauged catchment with two rainfall stations with a record of more than 10 years, and without any long records of discharge measurements. Rainfall in the catchment ranges from 550 mm yr⁻¹ in the lowlands (∼700 m) to 800 mm yr⁻¹ in the highlands (up to 2000 m) and is distributed over two rainy seasons. The short rainy season, locally known as “Vuli” runs from October to December, whereas the long rainy season, locally known as “Masika” runs from March to May. In 2004, the SSI Programme (Rockström et al., 2004) installed a hydrological monitoring network in the catchment, and has since been researching the hydrological implications of changing farmer management practices. The objectives of this study are to investigate the applicability of using hydrochemical tracers for hydrograph separation in a meso-scale semi-arid catchment in Tanzania.

Section snippets

Methodology

In this paper, chemical hydrograph separation for one relatively small flood event is described, using hydrochemical tracers including electrical conductivity (EC), dissolved silica (SiO₂), and major anions and cations. The flood event occurred on 9 November 2005, during the short rainy season in a sub-catchment of the Makanya catchment. The Vudee sub-catchment is monitored after the confluence of two rivers, upper-Vudee and Ndolwa (Fig. 2). At this location, discharge was recorded and samples

Results

During the event, the raingauges in upper-Vudee and Ndolwa observed 13.5 mm day⁻¹ and 7.9 mm day⁻¹, respectively. At the outlet in the valley, hardly any rainfall was recorded: 3.2 mm day⁻¹. Hourly rainfall records for this event were observed at this station, where a light drizzle started around 7 pm, continuing until 11 pm on 8 November. Rainfall in the upper part of the catchment was observed late afternoon. The runoff generated by this rainfall, started after midnight, reaching its peak at 2:30 am

Discussion

Hydrochemical tracers may be readily used for hydrograph separation, however hydrochemical tracers are a result of the processes that occur within the catchment, and therefore the assumption that the water quality of the fast runoff is equal to that of rainfall could be wrong and may introduce some uncertainty. This may also be the reason for the increase of some of the hydrochemical parameters at the beginning of the hydrograph. The effect on the water quality of surface runoff processes will

Conclusions

During the event of 9 November 2005, dominance of groundwater for flood formation was demonstrated by chemical hydrograph separations; over 95% was contributed by sub-surface runoff. This was confirmed by the complete lack of suspended sediments in the samples showing no erosion through surface runoff. Dissolved silica is a good tracer to distinguish between sub-surface and surface runoff, which was also found by Wels et al. (1991). Changes in selected chemical parameters can be used for the

Acknowledgements

The work reported here was undertaken as part of the Smallholder System Innovations in Integrated Watershed Management (SSI) Programme funded by the Netherlands Foundation for the Advancement of Tropical Research (WOTRO), the Swedish International Development Cooperation Agency (Sida), the Netherlands Directorate-General of Development Cooperation (DGIS), the International Water Management Institute (IWMI) and UNESCO-IHE Institute for Water Education. Thanks to the WaterMill (DGIS, The

References (13)

B. Ladouche et al.
Hydrograph separation using isotopic chemical and hydrological approaches (Strengbach catchment, France)
Journal of Hydrology
(2001)
V. Marc et al.
Investigation of the hydrological processes using chemical and isotopic tracers in a small Mediterranean forested catchment during autumn recharge
Journal of Hydrology
(2001)
U. Matsubayashi et al.
Hydrograph separation and flow analysis by specific electrical conductance of water
Journal of Hydrology
(1993)
O. Ribolzi et al.
Contribution of groundwater and overland flows to storm flow generation in a cultivated Mediterranean catchment. Quantification by natural chemical tracing
Journal of Hydrology
(2000)
J. Rockström et al.
A watershed approach to upgrade rainfed agriculture in water scarce regions through water system innovations: an integrated research initiative on water for food and rural livelihoods in balance with ecosystem functions
Physics and Chemistry of the Earth
(2004)
M.G. Sklash et al.
The role of groundwater in storm runoff
Journal of Hydrology
(1979)

There are more references available in the full text version of this article.

Cited by (47)

A four-component mixing model reveals snowpack melting as early as March during the 2019 hydrological year in the Upper Ganga Basin
2024, Journal of Hydrology
Ice and snow meltwater are important components of the glacier-fed Himalayan rivers that control the seasonal streamflow pattern. With the ongoing global warming, there remains significant uncertainty regarding how melting Himalayan glaciers will impact the water runoff. Here, we quantified the rainwater, ice meltwater, snow meltwater, and groundwater flow in the river using a δ¹⁸O, d, and Cl^- -based 4-component hydro-geochemical mixing model. At a weekly resolution, we measured the stable water isotopes (¹⁸O/¹⁶O and ²H/¹H, expressed as δ¹⁸O and δD), Cl^- and Na⁺ near the river Alaknanda mouth, in the Upper Ganga Basin over one year (2018–2019). The weekly scale time-series geochemical record captured the subtle variability of ice and snow meltwater flow in the Alaknanda River. The two most significant observation was (i) an onset of snow meltwater was observed as early as in March during the study period contrary to our present understanding that the onset of snow meltwater usually happens in April, and (ii) a significant increase in ice and snow meltwater flow during the peak monsoon season. The early onset of snow meltwater runoff in the Himalayan rivers could be an important concern as this could potentially reduce the flow of snow meltwater during the early summer months when water demand is highest due to intense agricultural activities in the mountainous region. Similarly, an increase in ice meltwater runoff in the monsoon season increases the risks of major flood events. Since weekly scale time-series data of ice meltwater, snow meltwater, rainwater, and groundwater flow in glacier-fed Himalayan rivers over the entire hydrological year was largely missing, this study provides an important dataset to better understand the year-long dependency of the Himalayan rivers on the glaciers critical to improving the accuracy and predictive power of hydrologic models.
Particularity of hydrological processes under heavy ablation based on environmental isotopes in transition zones between endorheic and exorheic basins
2023, Journal of Environmental Management
Drastic changes in the cryosphere have a significant impact on the quantity and formation process of water resources in the Qilian Mountains. The present study focused on quantitative evaluation of runoff components and runoff formation processes during strong ablation periods (August), in 2018, 2020, and 2021, in the transition zone between endorheic and exorheic basins in China, based on 1906 stable isotope samples. The results revealed that as the altitude decreased, the contribution of glacier and snow meltwater and permafrost water to runoff decreased, whereas that of the precipitation increased. Precipitation is a major source of river runoff in the Qilian Mountains. Notably, the runoff yield and concentration of rivers that were greatly affected by the cryosphere exhibited the following characteristics: (1) The altitude effect of stable isotopes was not significant and even showed a reverse trend in some rivers. (2) The processes of runoff yield and composition were relatively slow; as such, precipitation, glacier and snow meltwater, and supra-permafrost water were first transformed into groundwater and then supplied runoff to upstream mountainous region. (3) Finally, stable isotope composition in such rivers were similar to those in glaciers and snow meltwater, with small fluctuations. Therefore, the water sources of rivers affected by the cryosphere are more uncertain than those of rivers unaffected by the cryosphere. In future study, a prediction model of extreme precipitation and hydrological events will be developed, and a prediction technology for runoff formation and evolution in glacier snow and permafrost will be developed to integrate short-and long-term forecasts.
Hydrochemistry of waters in snowpacks, lakes and streams of Mt. Dagu, eastern of Tibet Plateau
2018, Science of the Total Environment
Citation Excerpt :
Besides, the isotopic composition of lake and stream waters behaved highly similar to snowpack samples, suggesting that lake waters at high altitude and stream waters at low altitude may originate from earlier meltwater of glacier. Hydrograph separation employing stable isotopes has been widely applied to identify the contributions of numerous water sources to runoff and evaluate streamflow generation mechanisms in a small spatial scale (Fritz et al., 1976; Mul et al., 2008). Based on Eqs. (3) and (4), the contribution proportion of meltwater and precipitation was calculated with the δ18O (similar result with δD).
There is little available information on hydrochemistry of waters from glacial source to downstream of glacierized catchments. Here we examine the water chemistry of the snowpacks, lakes and streams at eight sampling sites within glacial basin in Mt. Dagu, east Tibetan Plateau. An air mass trajectory model, correlation analysis, Gibbs model, Piper diagram and hydrograph separation analysis were utilized to investigate the characteristics and solutes sources of these waters. Generally, the TDS (Total dissolved solids; 7.54, 13.95 and 18.70 mg/L for snowpacks, lakes and streams respectively) and concentrations of main chemicals in all samples exhibited downstream trend from snowpacks to streams. Of the cations, Ca^2 + appeared with the highest concentration followed by K⁺ and Na⁺. Of the anions, HCO₃^– was most abundant, followed by Cl⁻, SO₄^2 − and NO₃⁻. For snowpack samples, the air masses arriving at the sampling sites were typically prevailing from the western Tibet Plateau, central Asia and the northern Mongolia plateau. The fine particulate matter in the Mt. Dagu snowpacks was most likely transported long range from three arid regions above-mentioned. High concentrations of SO₄^2 − and NH₄⁺ in snowpacks, with twice as much NH₄⁺ as SO₄^2 −, implying that the soluble part of the finer particles was transported as a form of ammonium sulfate. Rock weathering determined the ion components of the meltwater and the water could be classified as calcium and bicarbonate type based on the Piper diagram. The chemical contributions from glacier-snow meltwater were 20%―131% for lake and 5%―79% for stream, while the runoff contribution to lake varied from 65.4% to 84.9%, and 66.1% to 81.6% for stream. This study suggested that glacier-snow meltwater was the mainly runoff contributor to lake and stream water and that snowpack solutes derived from eolian additions exert a significant influence on lake and stream chemistry.
Evaluating relative merits of four baseflow separation methods in Eastern Australia
2017, Journal of Hydrology
Baseflow is the portion of streamflow that comes from shallow and deep subsurface flow, and is the key for catchment ecology and water resource management. This paper comprehensively evaluates four widely used non-tracer baseflow separation methods against tracer-based hydrograph separation for five Eastern Australian catchments. The four methods include United Kingdom Institute of Hydrology (UKIH) method and three digital filtering methods: Lyne–Hollick method, Chapman–Maxwell method and Eckhardt method. The first two filtering methods include a parameter of recession constant, and the last one has two parameters: the recession constant and the maximum baseflow index. We used an Automatic Baseflow Identification Technique (ABIT) to estimate the recession constant, which varies from 0.943 to 0.987 for the five catchments that is evidently higher than the default value of 0.925, and used the default Eckhardt and UKIH methods to estimate the maximum baseflow index, respectively. All modelling results are evaluated against the tracer-based hydrograph separation. Using the recession constant estimated from the ABIT method performs noticeably better than using the default parameter, indicated by the absolute bias reduced about 20% in average. For the two-parameter Eckhardt method, estimating the maximum baseflow index has larger effect on baseflow separation than estimating the recession constant. Compared to the different parameterisation schemes, the difference among the improved non-tracer methods is small. Using multiple passes into the Lyne–Hollick method can only slightly improve or deteriorate baseflow index estimates. Our results suggest that it is critical to get appropriate parameter(s) before applying the digital filtering methods.
Base flow separation: A comparison of analytical and mass balance methods
2016, Journal of Hydrology
Base flow is the ground water contribution to stream flow. Many activities, such as water resource management, calibrating hydrological and climate models, and studies of basin hydrology, require good estimates of base flow. The base flow component of stream flow is usually determined by separating a stream hydrograph into two components, base flow and runoff. Analytical methods, mathematical functions or algorithms used to calculate base flow directly from discharge, are the most widely used base flow separation methods and are often used without calibration to basin or gage-specific parameters other than basin area. In this study, six analytical methods are compared to a mass balance method, the conductivity mass-balance (CMB) method. The base flow index (BFI) values for 35 stream gages are obtained from each of the seven methods with each gage having at least two consecutive years of specific conductance data and 30 years of continuous discharge data. BFI is cumulative base flow divided by cumulative total discharge over the period of record of analysis. The BFI value is dimensionless, and always varies from 0 to 1. Areas of basins used in this study range from 27 km² to 68,117 km².
BFI was first determined for the uncalibrated analytical methods. The parameters of each analytical method were then calibrated to produce BFI values as close to the CMB derived BFI values as possible. One of the methods, the power function (aQ^b + cQ) method, is inherently calibrated and was not recalibrated. The uncalibrated analytical methods have an average correlation coefficient of 0.43 when compared to CMB-derived values, and an average correlation coefficient of 0.93 when calibrated with the CMB method. Once calibrated, the analytical methods can closely reproduce the base flow values of a mass balance method. Therefore, it is recommended that analytical methods be calibrated against tracer or mass balance methods.
Tracing water flow from sloping farmland to streams using oxygen-18 isotope to study a small agricultural catchment in southwest China
2013, Soil and Tillage Research
Citation Excerpt :
Genereux and Hooper (1998) concluded that the total stormflow runoff from rainstorm events is dominated by pre-event (“old”) water in a majority of small hillslope catchments. In addition, Saito (2000) and Mul et al. (2008) assumed that the isotopic concentration of subsurface runoff was identical to the value of pre-event water for use in hydrograph separation. In contrast, a few researchers have found that stormflow is dominated by event water (Bonell et al., 1998; Schellekens et al., 2004).
Knowledge of the age and origins of water and its movement from hillslopes to streams is important in mapping aquifers, conserving water supplies, determining water-use policy, and controlling pollution. In a sloping farmland-dominated agricultural catchment in southwest China, a stable isotope tracer (¹⁸O) combined with hydrometric measurements was tested to determine the rainwater movement from a hillslope to a stream during three storm events. The results showed that underflow in finely fractured rocks accounted for the highest proportion of the total flow on the slope, indicating that vertical flow was the predominant water flow pattern on the hillslope. The surface flow had basically the same δ¹⁸O value as the coinstantaneous rainfall, indicating that it was a Hortonian overland flow. The two end-member separation model showed that pre-event water accounted for a large proportion of the hillslope subsurface flows and streamflow especially during the initial stage. The first sample of underflow and interflow water always contained significant rainwater, indicating that macropore flow was prevalent in this Entisol. Event water dominated the peak streamflow and accounted for 70.4%, 95.6% and 80.4% of the peakflow amount during the three storm events. In general, the results suggested that pre-event water dominated the initial stage of stormflow generation and that event water dominated the peakflow. During the recession period, hillslope subsurface flows were the main source of the streamflow, as indicated by the similar volume-weighted δ¹⁸O value in subsurface flows and stream water. In future studies, hydrological processes under various land uses should be monitored synchronously to better understand the mechanisms of stormflow generation and transport regarding non-point source pollution in agricultural catchments.

View all citing articles on Scopus

View full text

Hydrograph separation using hydrochemical tracers in the Makanya catchment, Tanzania

Abstract

Introduction

Section snippets

Methodology

Results

Discussion

Conclusions

Acknowledgements

Journal of Hydrology

Journal of Hydrology

Journal of Hydrology

Journal of Hydrology

Physics and Chemistry of the Earth

Journal of Hydrology