2 Rationale for Optical Remote Sensing Using Satellite
Imagery
Satellite remote sensing has the potential to provide synoptic and frequent water
quality measurements of inland waters. Remote sensing satellites such as the
Landsat series have been collecting and archiving imagery regularly since the
early 1970s, which allows for the assessments of some historic water quality
information even on inland waters lacking historical ground-based data. Satellite
systems planned for launch in the next few years will allow better characterization
of inland water quality on regional-to-global scales.
Optical remote sensing (ORS) using satellite imagery can be used to measure
water quality of inland, marine, and coastal waters. Although there are many
similarities between ORS applied to inland waters and ORS applied to marine
systems, there also are profound differences. For example, spatial resolution
requirements are much lower for the broad expanses of the oceans and most coastal
areas than are needed for small inland water bodies. Several generations of satellite
sensors acquire images with large pixel sizes (~0.3–1 km) that provide adequate
spatial resolution for oceanic and most coastal studies but are too coarse for small
inland water bodies. For perspective, the smallest water body that can be measured
by a satellite sensor with a pixel size of 1 km is ~1,000 ha [1]. The spatial resolution
of Landsat satellites, 30 m, generally allows measurements on water bodies larger
than ~4 hectares (ha). As pixel size increases, the likelihood decreases that an image
will have at least one pixel (preferably four or more) focused solely on open water
and not affected by terrestrial and shallow near-shore areas. The smaller pixel size
also allows for better characterization of bays and narrow portions of complex lake
systems
A second difference relates to the optical complexity of inland waters. Remote
sensing scientists focusing on marine systems are able to use increasingly sophisticated instrumentation such as the Moderate Resolution Imaging Spectroradiometer
(MODIS) aboard the Aqua and Terra satellites to develop analytical and semianalytical algorithms that retrieve chlorophyll levels from the oceans, and this has
become a routine, global-scale operation [2, 3]. Remote sensing scientists focusing
on inland waters have had to develop procedures primarily using other satellites like
Landsat, which have adequate spatial resolution but at the same time have critical
deficiencies in spectral and temporal resolution. The blue and green spectral bands
used to retrieve chlorophyll levels from oceanic waters are not so useful for such
purposes in optically complex inland waters [4–7]. These deficiencies have limited
development of retrieval algorithms for inland water quality variables by satellite
imagery mostly to empirical and semiempirical (described in Sect. 3.1) approaches.
More sophisticated ground-based and aircraft-mounted spectroradiometers also
have been used in recent years to advance the science of inland water ORS.
In summary, the requirements for spatial resolution, most effective spectral bands,
ability to use analytical (versus empirical) approaches, and ranges of interest for
water quality variables like chlorophyll and CDOM are different between inland
Remote Sensing for Regional Lake Water Quality Assessment: Capabilities and. . .
113
Imagery
Satellite remote sensing has the potential to provide synoptic and frequent water
quality measurements of inland waters. Remote sensing satellites such as the
Landsat series have been collecting and archiving imagery regularly since the
early 1970s, which allows for the assessments of some historic water quality
information even on inland waters lacking historical ground-based data. Satellite
systems planned for launch in the next few years will allow better characterization
of inland water quality on regional-to-global scales.
Optical remote sensing (ORS) using satellite imagery can be used to measure
water quality of inland, marine, and coastal waters. Although there are many
similarities between ORS applied to inland waters and ORS applied to marine
systems, there also are profound differences. For example, spatial resolution
requirements are much lower for the broad expanses of the oceans and most coastal
areas than are needed for small inland water bodies. Several generations of satellite
sensors acquire images with large pixel sizes (~0.3–1 km) that provide adequate
spatial resolution for oceanic and most coastal studies but are too coarse for small
inland water bodies. For perspective, the smallest water body that can be measured
by a satellite sensor with a pixel size of 1 km is ~1,000 ha [1]. The spatial resolution
of Landsat satellites, 30 m, generally allows measurements on water bodies larger
than ~4 hectares (ha). As pixel size increases, the likelihood decreases that an image
will have at least one pixel (preferably four or more) focused solely on open water
and not affected by terrestrial and shallow near-shore areas. The smaller pixel size
also allows for better characterization of bays and narrow portions of complex lake
systems
A second difference relates to the optical complexity of inland waters. Remote
sensing scientists focusing on marine systems are able to use increasingly sophisticated instrumentation such as the Moderate Resolution Imaging Spectroradiometer
(MODIS) aboard the Aqua and Terra satellites to develop analytical and semianalytical algorithms that retrieve chlorophyll levels from the oceans, and this has
become a routine, global-scale operation [2, 3]. Remote sensing scientists focusing
on inland waters have had to develop procedures primarily using other satellites like
Landsat, which have adequate spatial resolution but at the same time have critical
deficiencies in spectral and temporal resolution. The blue and green spectral bands
used to retrieve chlorophyll levels from oceanic waters are not so useful for such
purposes in optically complex inland waters [4–7]. These deficiencies have limited
development of retrieval algorithms for inland water quality variables by satellite
imagery mostly to empirical and semiempirical (described in Sect. 3.1) approaches.
More sophisticated ground-based and aircraft-mounted spectroradiometers also
have been used in recent years to advance the science of inland water ORS.
In summary, the requirements for spatial resolution, most effective spectral bands,
ability to use analytical (versus empirical) approaches, and ranges of interest for
water quality variables like chlorophyll and CDOM are different between inland
Remote Sensing for Regional Lake Water Quality Assessment: Capabilities and. . .
113