Geography Reference
In-Depth Information
remains as a dominant application of remote sensing.
Here we find applications of LiDAR, colour and multi-
spectral data aimed at identifying the composition and
land-uses of the riparian zones along with their tem-
poral dynamics. Traditional image classification of these
datasets remains the principal method. However, a few
papers did mention newly developed object-based classi-
fication methods. We also find that the scale of the studies
in this category varies quite widely from studies focusing
on bankside vegetation a few meters or tens of meters
away from the channel to studies examining the entire
catchment of large rivers.
Studies of river water chemistry and suspended sedi-
ment concentrations (SSC) are well established in oceanic
sciences. They are also well established in large river
science with some early work taking advantage of the
Landsat program (Aranuvachapun and Walling, 1988).
However, in the context of the smaller, so-called 'nor-
mal', rivers which are the focus of this topic, remote
sensing of water quality publications is scarce. Within our
search results, water-quality papers were dominated with
estuarine and large river studies at the interface between
oceanic and fluvial sciences. The rationale behind most
of these studies is to replace expensive and labour inten-
sive ground-based field monitoring with multispectral or
hyperspectral remote sensing data in order to perform
what is in essence 'remote spectroscopy'. The key focus
is the study of river plumes in terms of sediment load
and pollution load. The parameters which are directly
measured in these studies are turbidity (i.e. water clar-
ity) and organic matter concentrations (i.e. presence of
chlorophyll). These metrics can then be used as proxies
for other parameters such as pollution load and salin-
ity. The most commonly used sensors are the ETM
+
(Landsat), MODIS and the Advanced Land Imager (ALI)
which is amultispectral sensor mounted onNASA's Earth
Observation-1 (EO-1) satellite.
1.3.4 Spatial andTemporalResolutions
Finally, the abstracts were used to examine the range
of temporal and spatial resolutions in use within our
abstract database (Figure 1.8a). Most of the contributions
are based on spatial resolutions of 10-50m confirming
the use of satellite imagery. Coarser resolutions are less
frequent. Metric and sub-metric resolution mainly based
on airborne imagery are also very common, reaching 25%
of papers. Ground-based remote sensing (here we assume
decimetric or centimetric resolutions even if not speci-
fied in the abstract) is also a field which is well explored
within 10% of the contributions. When combining top-
ical areas and the spatial resolutions, a
2 test shows
they are dependant (p < 0.0001) (Table 1.3). Discharge
and fishmonitoring are based on ground remote-sensing,
whereas DEM and bathymetry use very high resolution
(
χ
1m) data. Riparian features and land-use mapping
also used very high resolution (1m) data mainly based
>
0
5
10
15
20
25
Daily
Day
Month
Season
Year
Yearly
Annual
Inter-annual
Decadal
Decade
Decennal
Century
History
Multi-date
Multi-temporal
Archive
No
Indications
Review
>50 m
10-50 m
1-10 m
1
<1 m
Ground
0
10
20
30
(a)
(b)
Figure 1.8 Frequency of terms within the abstracts of the 200 manuscripts (in % of the studied manuscripts) dealing with (a) spatial
and (b) temporal framework.
Search WWH ::




Custom Search