1. The Humber Watershed
2. The Impact of Urbanization on Watershed Hydrology
3. Determining Land Cover via Remote Sensing
4. Determining Catchment Areas using Digital Elevation
Models
5. The SCS Method of Runoff Modelling
1. The Humber Watershed
The Humber Watershed covers 908 sq-km and is home to 485 000 people. It is
the largest of the 9 watersheds in the Metropolitan Toronto and Region
Conservation Authority (MTRCA) jurisdiction. Groundwater originating in
the Humber basin feeds more than 100 tributaries (MTRCA, 1995).
Within the Humber there are five distinct physiographic regions. These
play a strong role, along with climate, in determining the functioning
of the hydrologic cycle (and in determining habitat for flora and
fauna). The five regions are (1) the Niagara Escarpment (2) the Oak
Ridges Moraine (3) Peel and Iroquois Plains (4) Valley and Stream
Corridors, and (5) Waterbodies (HWTFNHS 1996).
The Watershed is a dynamic system. The greatest changes to the Watershed
since the last glaciation have occurred in the last 200 years, with
human impacts.
Impacts in the Oak Ridges Moraine are of particular
concern with regard to the hydrologic cycle. The Moraine, encompassing
27% of the area of the Watershed (HWTFNHS 1996), is comprised of coarse,
well-drained soils. The coarse soils allow meltwater and rainfall to
soak into the ground, feeding aquifers. These aquifers, in turn,
discharge cold, and relatively clean, water, feeding the headwaters of
the East Humber and Main Humber rivers, as well as their tributaries
(HWTFNHS 1996). The aquifers also provide the principal water supply for
many communities and rural residents. Fortunately, the Moraine is
largely in a relatively undisturbed
state. However, development in the Moraine threatens the long-term
quality and quantity of local and regional water resources.
While the Oak Ridges Moraine is vital, urbanization in any part of the
Watershed impacts the hydrology of the basin. The theoretical background
with regard to the impact of urbanization on the Watershed is discussed
in the next section.
Resources:
Go to top
Water is a vital component of the natural environment, influenced by
climate and land, and sustaining life. This is best understood in terms
of the hydrologic cycle. The hydrologic cycle
can also be seen as a system diagram.
Urbanization impacts the hydrologic cycle in two key ways. Through (1) increasing population density, and (2) increasing
building density. This can be represented graphically. The result is a more complicated hydrologic cycle. There are two major effects of urbanisation on water quantity, (1) increases
in runoff volumes, and (2) increases in peak flow rates. Urbanization
impacts water quality in that it results in increased loads of metals
and organic contaminants.
The impact of urbanization on watershed hydrology has been explored in a
number of studies. Cook et al. (1985) completed a study of the Speedvale Experimental Basin (a 210 ha watershed on
the outskirts of Guelph, Ontario). A relatively extensive hydrologic
data base of precipitation, streamflow, soil moisture, and groundwater
was assembled both for the pre-urbanization period (1966-74) and the
period of development (1975-82). The study area was used for
agricultural purposes in the first stage. In the second period, 155 ha
was serviced for light industrial and commercial development, which
involved the installation of a stormwater conveyance system.
Urbanization resulted in an increase in impervious area from 1% to 33%
of the basin.
Cook et al. (1985) found that "...changes in land use coincided with changes in
volumetric and time distribution aspects of hydrologic response. Changes
in the response include:
(i) an increase in the mean annual runoff
coefficient by a factor of 1.5;
Boyd et al. (1993) completed a similar study, except that they
contrasted pervious and impervious surfaces rather than urban and rural
areas. They studied runoff from pervious versus impervious
surfaces in 26 urban catchments in 12 countries. Rainfall and runoff
depths were examined for 763 storm events. It was found that the
effective impervious area remained constant for all storm sizes. This is
significant in that it indicates that impervious surfaces will cause the
vast majority of precipitation to runoff, with little other system response.
Pervious runoff, by contrast, was determined by the depth of rain in the
storm for rainfalls less than 50 mm, while rainfall intensity and
antecedent wetness were also significant with regard to runoff from
larger storms.
Stephenson (1994) studied two topographically similar, and adjacent,
catchments near Johannesburg, South Africa over a five year period. One of the catchments was
suburban while the other was natural grassland. Surface runoff from the
undeveloped and suburban catchments was 4% and 15% respectively.
Evapotranspiration was 63% for both catchments. The study concluded that
suburban development increased surface runoff volume by a factor of four.
It has long been recognized that urban runoff contains a diverse range
of pollutants, which can occur in concentrations that are harmful to the
ecosystem. A profile of impacts was developed by the U.S. Environmental
Protection Agency in 1992, which found that "between one-third and
two-thirds of all designated use impairment for...streams was a result of
agricultural and urban runoff" (Debo & Reese, 1995: 673).
Although there is a dearth of hard data, it is generally acknowledged
that urban runoff contains significant levels of pollution and is a
major source of receiving water pollution. The Humber Watershed Task
Force has recognized that industrial land use generally contributes
relatively high concentrations of metals, while agricultural and
residential land use tends to result in high organic concentrations in
runoff (1996).
Resources:
Go to top
3. Determining Land Cover via Remote Sensing
As explained in the above discussion, both theory and empirical studies
indicate that changes in land use resulting from urbanisation are linked
to changes in resulting runoff volumes and rates.
In order to track changes in land use over time, a data source for land
use is required that describes land use in a fashion that is both
sufficiently detailed and easily processed for analytical purposes.
Multispectral remotely sensed images are an ideal data source for this
application.
Remote sensing is defined as the measurement, from a distance, of the
spectral futures of the earth's surface (Mather 1987). Data collected over time can be
used to determine changing land cover/use.
Once a land use classification map is derived from the multi-spectral images,
land uses can be associated with runoff production through some form of
hydrological modelling.
Resources:
Go to top
In order to associate the land use classifications derived from multispectral remote
sensed images with the stream quality and quantity data taken at a point on the
stream network, the catchment that drains the network to that point must be
determined. By defining the catchment associated with a given stream gauge, two
objectives are fulfilled: 1) The area of the catchment is known, allowing volumetric
discharges to be converted to depth equivalent values for easy computation and
comparison with precipitation. 2) The spatial region that drains the landscape to
the stream quality/quantity station is explicitly identified, allowing the
computation of descriptions of land cover that are directly associated with the
stream station.
Determining a catchment area from a raster digital elevation model (DEM) involves the
invoking of the most basic hydrologic rule, namely that water flows downhill. A
raster DEM conceptualises the landscape in terms of a large grid of adjacent
squares that each have an associated value of altitude. Applying the downhill flow
rule to this model of the landscape results in water from a square (pixel) will flow
to the adjacent square that has the lowest elevation. Using this principle, a
drainage direction map is produced that indicates in which direction any given pixel
will drain. By specifying the location of the stream monitoring station, the
associated catchment can be calculated from the drainage direction map, by working
uphill from the location, finding every pixel that eventually drains to the location.
The above description is a somewhat simplified expression of what is involved
algorithmically. In a practical sense, these issues are simplified for the user of
GRASS as well, as the procedure is easily implemented through the r.watershed
(drainage direction generation) and r.water.outlet (catchment for a stream
outlet) programs.
Resources:
Go to Top
The Soil Conservation Service (1972) developed a method for abstracting
runoff from rainfall based on the land use and soil type characteristics
of the catchment. Based on these two factors, an appropriate SCS curve
number (CN) can be selected from a table of values (determined through
empirical analysis of several catchments). Also, a further table
provides CNs for urbanized areas. Once a CN is chosen, resulting runoff
can be determined from precipitation using an equation or diagrammatic
form of the SCS model. Using this procedure, simulated runoffs are
calculated and compared to actual runoffs. The model is calibrated by
modifying the CN value until the model predicts runoff accurately.
Resources:
Go to Top
2. The Impact of Urbanization on Watershed Hydrology
(ii) an increase in the instantaneous
discharge by a factor of almost 3.0;
(iii) a change in the time of
annual peak flow from occurring solely in the spring runoff period to
occurring throughout the various seasons;
(iv) a change in the seasonal
pattern of monthly and storm event runoff coefficients with the greatest
change observed in the summer and lesser changes observed in the other
seasons;
(v) a 3 fold reduction in unit hydrograph lag time; and
(vi) a
3.5 fold increase in unit hydrograph peak flow".
4. Determining Catchment Areas using Digital Elevation Models
5. The SCS Method of Runoff Modelling
![[UP]](../images/up_but.gif)
