50
Geotechnical News • December 2015
WASTE GEOTECHNICS
Prediction of rainfall runoff for soil cover systems:
A laboratory approach
Ahlam Abdulnabi
Introduction
For the past few decades, soil cover
systems have proven efficiency as
engineered barriers to manage hazard-
ous mine waste. The purpose of these
covers can vary, but usually aims to
restrict water and oxygen contact with
sulphide minerals in waste rock and
tailings, thus preventing the onset of
Acid Rock Drainage.
The different types of soil covers
and the factors considered in their
design are summarized in O’Kane
and Wels (2003) and the GARDGuide
(The International Network for Acid
Prevention (INAP) 2009). These types
can be broadly classified into water
covers and dry covers, which can be
subcategorized as ‘conventional low
hydraulic conductivity’ covers, ‘capil-
lary barrier’ covers, and ‘store and
release’ covers.
Generally, the design of dry covers is
governed by the amount of net infiltra-
tion into the system. Prediction of
said net infiltration requires a detailed
soil-atmosphere modelling using site-
specific infiltration models. Attaining
accurate results from these models
involve proper calibration using water
balance equations. Thereby, necessi-
tate the accessibility to runoff predic-
tions, since surface-runoff can be the
largest component of the water budget
that directly influences the amount of
net infiltration.
Prediction of rainfall runoff is particu-
larly crucial not only for the design,
but also for the longevity assessment
of dry cover systems. Current meth-
ods of rainfall runoff prediction entail
either complex modelling of infiltra-
tion at the point scale (Green and
Ampt 1911; Horton 1939; Philip 1957;
Mein and Larsen 1978); or require
estimates prone to inevitable temporal
and spatial variations in soil properties
and rainfall events at the watershed
scale (Schmocker-Fackel et al. 2007,
Benson, 2010 and Jubinville 2013).
Reliable models for predicting surface
runoff at the field scale based on
quantifiable soil properties seem, by
and large, scarce, and require improve-
ments especially when it comes to
taking different initial state or anteced-
ent moisture conditions of the soil into
account (Abdulnabi 2015).
Since nothing beats repeatable verifi-
able observations, a laboratory-testing
program was established to address
this need for a reliable model to
predict rainfall runoff for soil cover
systems. The program investigates
the correlation between laboratory-
induced rainfall of different intensities,
and the subsequent runoff response
in both ‘low hydraulic conductivity’
and ‘capillary barrier’ soil covers.
The primary focus of that program is
to identify the appropriate variables
that control runoff generation for both
saturated and unsaturated state.
Description of the laboratory
program
The laboratory program was con-
ducted using a specially designed
rainfall simulator apparatus. The main
components of the apparatus were a
water circulation system, a spraying
system, and a flume to accommodate
the soil and the measuring devices.
The water circulation system consisted
of a water reservoir equipped with
a submersible constant-rate pump
to direct the water to the spraying
system.
The spraying system comprised of a
number of nozzles of different orifices.
Each set of nozzles produced different
rainfall intensity. The most appropriate
type of nozzles for this study was the
one that provides an even distribution
of medium-sized raindrops throughout
a rectangular spray pattern. The height
of the spraying arm was obtained by
iteration trials to achieve the correct
spray pattern, the optimum rainfall
coverage of the plot, and the maxi-
mum uniformity of simulated rainfall.
Similarly, the spacing between the
nozzles was also obtained to eliminate
overlapping of raindrops and to ensure
concordant coverage of the plot.
The soil was accommodated inside a
transparent flume to enable observing
the wetting-fronts propagation as tests
progressed. The dimensions of the
box were 900mm in length, 300mm
in width, and 350mm in height. The
flume had a runoff collection outlet
at the top, and a one-inch drainage
opening at the toe. Measuring devices
included Time Domain Reflectometry
(TDR) probes and Tensiometers for
measuring volumetric moisture con-
tent and matric suction, respectively.
Instruments were distributed evenly
at two elevations in the flume. A view
of the overall setup is illustrated in
Figure 1.
