Mapping the Extent of Flooding - Remote Sensing Application -
Mapping the Extent of Flooding

This kind of information was a factor in predicting the major flood in the northern Midwest, late in the Spring of 1997. Satellite imaging gave a critical look at the great floods on the Red River in North Dakota and Manitoba (Canada) that inundated Grand Forks, Fargo, and other towns along the state's border with Minnesota.

Spring flooding is frequent in parts of the Mississippi River basin. A hundred-year flood, i.e., largest expected statistically in a 100-yr span, resulted from snow melt and rain in late March of 1973. This Landsat-1 subimage (with an earlier pre-flood view), captured the extent of flooding on a cloud-free day, showing St. Louis, Missouri (protected from downtown flooding), and the flood plains of the Mississippi, Missouri (joined at A), and Illinois (at B) rivers:

 Left view: Landsat-1 false color composite of St. Louis, MO, and the Mississippi, Missouri, and Illinois Rivers in normal flow; right view: same area experiencing a 100-year flood in March, 1973.

Twenty years later, the Midwest again flooded, worse than before, in fact many places recorded the highest flood levels in their history. After several months of excessive rain that saturated the soil, because of a blocking high pressure system that kept the jet stream relatively stationary, in late July and August of 1993, water levels rose well above flood stage. Areas hardest hit were from Iowa to southern Illinois. Levees broke, inundating tens of thousands of acres. The '93 flood, said to be largest ever on the Mississippi, became the costliest in U.S. history (some estimates approach $15 billion). Satellite imaging played a key role in getting a number of good images of the flooded area.

Once again we examine the lowlands northwest of St. Louis. First is this Landsat image:

Satellite image of flooding in the St. Louis region.

The next image was taken by Shuttle astronauts, using SIR-C radar. (Note that the orientation has west near the top.) Below that is an image of merged JERS-1 radar and a SPOT 3-band composite, which offers considerable detail (notice how farmlands show through the water).

Color SIR-C radar image of a flood in the lowlands northwest of St. Louis, Missouri, August 1993.
Merged color JERS-1/SPOT-3 image of a flood in the lowlands northwest of St. Louis, Missouri, August 1993.

The flooding of 1993 affected many rivers. Floods occur as early as June and as late as October (see this NOAA website for an historical review). The next three photos show the flooding at St. Louis (on August 1 the flood peaked at the waterfront at 49.6 ft, 19 feet above flood stage and 6 feet higher than during the 1973 flood shown above), Jefferson City, MO, and Alton, IL:

Flooding at the St. Louis Arch.
Flooding of the Missouri at Jefferson City.
Flooding near Alton, IL.

Up river during the same 1993 flood, the SAR radar on ERS-2 rendered the flood in mostly black tones in this scene near Dubuque in southeast Iowa:

Flooded land in eastern Iowa, inundated in the 1993 flood, as imaged by the ERS-2 radar.

The effects of the 1993 persisted in the lowlands near St. Louis well into 1994, as indicated by this March 1994 photo taken from the Space Shuttle:

Flooding near St. Louis.

Flooding can occur anywhere on all continents except the Antarctic. The next image is a Landsat-1 subscene (February 6, 1974) of the Barcoo River in Queensland/South Australia, flooded by Fall rains. The floodwaters have spread greater than 50 km (31 mi) wide in these low-lying plains, with low rolling hills.

Landsat-1 subscene image of a flood of the Barcoo River in Queensland, South Australia, February 6 1974.

The Yangtse River in China underwent a major flood in August of 1998. Millions were driven from their lowlands homes. This Radarsat image shows the flooded lands near Wuhan:

Radarsat image of floods along the Yangtse River.

In this unusual image, an ERS-1 radar image taken during June of 1993 is joined with an ERS-2 radar image taken on August 1, 1998, providing a multitemporal or change detection rendition. Both blue and red associate with flood waters.

A multitemporal radar image using a ERS-1 June, 1993 view of an area along the Yangtse River registered with a ERS-2 August 1, 1998 scene of the same area.

Radar is especially powerful in recognizing floodwater extent since the low reflectivity from water makes it appear dark. Floods can be beneficial as well as harmful. These two Radarsat images show annual flooding in the Mekong River basin in Cambodia. Waters from the monsoon rains collect in the basin and replenish the ubiquitous rice paddies that provide the main food supply for the Vietnamese and Cambodians.

Radarsat images of the wet season flooding compared to dry season decreased water distribution (some areas do have standing water) in the Mekong River basin.

The pair of MODIS images below show parts of Namibia, Botswana, Zambia and Zimbabwe in which the rivers are normal and then in flood:

MODIS image of parts of southwest Africa.
Same region as above, with flooding.

Finally, check this map showing areas with high soil moisture, a condition that bespeaks of water saturation from earlier heavy rainfall or from previous flooding. This soil wetness map shows much of Asia; maps of all the continents plotting this parameter are made by NOAA/NESDIS.

 A Soil Wetness Index map for June 2000 made from NOAA satellite observations over much of Asia.

This lengthy Section 14 purports to convey that the principal use of remote sensing remains surveillance of weather systems and oceans on local-to-global scales. We report this because of the widespread occurrence of water on the Earth's surface (even greater than the 70+% ocean surface area, stated on page 14-1, if we include the Antarctic ice [which stores more than 80% of the world's {frozen} fresh water] and Greenland.

At this point in the Tutorial, we have examined most of the specialized modes of remote sensing (defined by the electromagnetic spectral regions we can use), the spacecraft systems that mount the sensors, and the numerous applications to which these sensors have contributed. In the next Section (15), on Geographic Information Systems (GIS), we look at some systematic ways to integrate remote sensing data into organization, correlation, interpretation, and management of geographically-referenced information. Then, in Section 16, we look ahead to some of the current or recent remote sensing programs, in which individual satellites will simultaneously make meteorological, oceanographic, land surface, and biologic observations to present a unified picture of Earth as a System.