310
A.M.J. Meijerink
(see discussion of Fig.14.4), localized high drainage densities and transfonnation
of river types, such as the change of ephemeral wide sandy river bed, to narrow,
sinuous perennial rivulets, Fig.14.3b.
Infonnation on piezometric heads under confmed or semi-confined conditions in
regional hydrogeologic studies, is often absent or available only for productive
aquifers. Areas with expected differences between phreatic heads and piezometric
heads may be estimated by hydrogeologic reasoning, but actual observations are
required and remote sensing may contribute only in an indirect manner. It may be
remarked here that reconstruction of the groundwater surface is a manual
operation, unless so many data points are available that one of the interpolation
methods can be meaningfully employed (Meijerink et al. 1994).
Two limitations should be mentioned. There are areas, often cultivated, where
the groundwater table is at shallow depths, say, I to 3 meters, and where no
evident recognition features can be detected on the imagery. This is true for most
of the area ofFig.14.1. The other problem is to judge, in upland areas, whether the
interpreted shallow groundwater table is perched and possibly seasonal or whether
it is the surface of the coherent groundwater body. Merging of image
interpretations with hydrogeologic field data may resolve the issue.
14.2.3 Flow systems
The theory of groundwater flow systems must be credited to Toth (1962) and it
provides an "in-depth"- background for hydrogeologic image interpretations. It has
been estimated that for groundwater basins 5 to 30 % of the surface area consist of
groundwater discharge - or exfiltration- areas. The region with the sinuous small
rivers in Fig.14.3b is such a discharge area, and the area with the grasslands in the
centre of the colour image (Colour Plate 14.A).
Figurel4.4 is included to illustrate how remote sensing can be used in
developing a conceptual hydrogeologic model including flow systems in a data
scarce area (Western Province, Zambia). The image shows the floodplains of the
Zambezi River and a tributary and the gently undulating interfluve of the Kalahari
Sands. The image is composed of superimposed NDVI MSS images of three
different seasons. Hence, whitish tones correspond to pennanent green vegetation.
The area has semi-circular depressions called Dambo's. In some, intersection with
the groundwater table occurs, as is indicated by perennial water. This infonnation,
together with well data, has been used for the estimation of the groundwater
contours. The pennanent green vegetation along the fringe of the floodplain where
the topography rises, has been interpreted as an exfiltration area of a flow system.
Two-dimensional mod~lling supports the presence of discharge zones at those
places. These results are used for setting up a groundwater model, whereby the
chief purpose is, at the initial stage, to guide further exploration.
A.M.J. Meijerink
(see discussion of Fig.14.4), localized high drainage densities and transfonnation
of river types, such as the change of ephemeral wide sandy river bed, to narrow,
sinuous perennial rivulets, Fig.14.3b.
Infonnation on piezometric heads under confmed or semi-confined conditions in
regional hydrogeologic studies, is often absent or available only for productive
aquifers. Areas with expected differences between phreatic heads and piezometric
heads may be estimated by hydrogeologic reasoning, but actual observations are
required and remote sensing may contribute only in an indirect manner. It may be
remarked here that reconstruction of the groundwater surface is a manual
operation, unless so many data points are available that one of the interpolation
methods can be meaningfully employed (Meijerink et al. 1994).
Two limitations should be mentioned. There are areas, often cultivated, where
the groundwater table is at shallow depths, say, I to 3 meters, and where no
evident recognition features can be detected on the imagery. This is true for most
of the area ofFig.14.1. The other problem is to judge, in upland areas, whether the
interpreted shallow groundwater table is perched and possibly seasonal or whether
it is the surface of the coherent groundwater body. Merging of image
interpretations with hydrogeologic field data may resolve the issue.
14.2.3 Flow systems
The theory of groundwater flow systems must be credited to Toth (1962) and it
provides an "in-depth"- background for hydrogeologic image interpretations. It has
been estimated that for groundwater basins 5 to 30 % of the surface area consist of
groundwater discharge - or exfiltration- areas. The region with the sinuous small
rivers in Fig.14.3b is such a discharge area, and the area with the grasslands in the
centre of the colour image (Colour Plate 14.A).
Figurel4.4 is included to illustrate how remote sensing can be used in
developing a conceptual hydrogeologic model including flow systems in a data
scarce area (Western Province, Zambia). The image shows the floodplains of the
Zambezi River and a tributary and the gently undulating interfluve of the Kalahari
Sands. The image is composed of superimposed NDVI MSS images of three
different seasons. Hence, whitish tones correspond to pennanent green vegetation.
The area has semi-circular depressions called Dambo's. In some, intersection with
the groundwater table occurs, as is indicated by perennial water. This infonnation,
together with well data, has been used for the estimation of the groundwater
contours. The pennanent green vegetation along the fringe of the floodplain where
the topography rises, has been interpreted as an exfiltration area of a flow system.
Two-dimensional mod~lling supports the presence of discharge zones at those
places. These results are used for setting up a groundwater model, whereby the
chief purpose is, at the initial stage, to guide further exploration.