SUBSURFACE SOIL MOISTURE WITH PASSIVE MICROWAVE RADIOMETRY

Microwave radiometry, or passive microwave remote sensing, is one of the methods used for remote observations of the environment. It is based on the measurement of natural electromagnetic radiation of objects in the microwave range of wavelengths. Between the wavelengths of 2 to 30 cm, the Earth's radiation is primarily a function of the free water content in soil. EarthData Technologies, working with personnel from the Russian Academy of Science, has developed the ability to collect geo-referenced airborne microwave data with a relatively small ground sample distance. In November, 1997, the US Army Corps of Engineers conducted a flight along the Mississippi River near Cairo Illinois to asses the feasibility of using this technique to map the spatial variation of subsurface soil moisture across the river basin, and in particular, along the federal levee system. The data was merged with the Corps' River Engineering, Environmental and GIS (REEGIS) data base layers containing the streets, highways, and the federal levee system. The resulting data successfully presented areas of subsurface moisture that are known to correspond to water intrusion, under the levee, during flood events.

The scope of this project was to collect and process passive microwave data over several sites along the Mississippi River. The data was collected to establish a baseline of the spatial variation of subsurface soil moisture along the river levees. The overall purpose of the project was to: 1) conduct the survey to measure soil moisture; 2) to document the results showing the ability of the system to detect subsurface moisture variations, and; 3) to assess the feasibility of expanding this technology throughout the river system to locate and quantify levee saturation or levee under seepage during river flooding.

Microwave radiometry, or passive microwave remote sensing, is one of the radio-physical methods used for remote observations of the environment. It is based on measurements of the natural electromagnetic radiation of objects in the millimeter to decimeter range of wavelengths [Ref 1-4].

Investigations of water and land surfaces may occur in the 2 to 30 cm spectral band. Between these bands, the Earth's radiation is primarily a function of the free water content in soil. The response within the 2 to 30 cm band is also influenced by other parameters, such as depth to a shallow water table, vegetation biomass, salinity and temperature of open water, where the sensitivity is a function of the wavelength [Ref 1-7].

The measure of the intensity of radiation in the microwave band is referred to as a brightness temperature, T_b , which is a product of emissivity, k , and thermodynamic temperature, T_e, within the effectively emitting layer (skin-depth) of the object.

T_b = k T_e

The thermodynamic and brightness temperatures are measured in Kelvins where,

T(K) = t(C) + 273.

Within the 2 to 30 cm band, for t=10-30 ^oC, the characteristics of several surface types are shown in Table 1.

Table 2 shows the sensitivity of radiation in the X-band (2-3 cm) and L-band (18-30 cm) to the changes in free water content in soil, soil density, salinity and temperature of the soil surface [Ref 1, 2]. These data show that the main parameter affecting the intensity of radiation, independent of spectral band, is the soil moisture. Based on this sensitivity it is feasible to estimate the value of soil moisture without a priori data on the soil parameters.

It has also been seen that even for rather high values of biomass, up to 2 to 3 kg/m² , the plant canopy is still transparent in the decimeter wavelength range.

There are several types of instrumentation that can be used for "environmental" investigations. For examination of soil type, or surface moisture, the ones that come to mind are Thermal Infrared cameras, active radar systems, and the microwave radiometer used for this project. The radar systems, including Interferrometric Synthetic Aperture Radar (IFSAR), share the same frequency band as the passive radiometer. Therefore they have similar wave propagation characteristics, although the radar systems are actively transmitting and measuring the return values. The radar / IFSAR systems rely on surface texture to deduce surface boundary conditions. The passive system used on this project only records the naturally emitted radiation, which is a direct function of the soil moisture content within the first several meters. From a given height above the ground surface the radar systems have a much smaller ground sample distance, or pixel size, than the passive system. To compensate, the passive system can be flown on light aircraft at very low altitudes, as low as 500 feet above mean terrain.

Thermal Infrared sensors (TIR) and the passive radiometer both measure natural radiation from a thermodynamic process. However TIR are sensitive to the temperature variations which take place within the surface layer of the object being imaged. This temperature variation could be the result of many things, such as warm water, or steam leaking from underground pipes, poor insulation properties on buildings, effluent from a manufacturing process, etc. Passive microwave radiometers, on the other hand, measure the natural radiation that is a direct function of soil moisture in the first couple of meters of the earth's surface. TIR sensors are further influenced by cloud, fog, smoke, solar radiation and other similar atmospheric conditions.

The passive microwave technology has been under development at the Institute of Radioengineering and Electronics, Russian Academy of Sciences since the 1950's. The systems have been developed in many configurations including single beam, multi-beam, scanning, etc., and have flown on a wide array of large and small fixed wing aircraft and helicopters. In 1968 the Cosmos 243 spacecraft was launched, equipped with four (4) single beam microwave radiometers operating in the wavelengths of 0.8 cm, 1.35 cm, 3.4 cm and 8.5 cm. In April 1996 a radiometer was included on the payload of the Piroda spacecraft that has been attached to the Mir Space Station.

In support of the USACE, EarthData utilized one of its aircraft, a Piper Navajo Chieftain. This aircraft is fully equipped with avionics for safe mission operation throughout North and South America, including communications gear supporting missions over military ranges. Further, the craft utilizes automated flight navigation based on GPS, and predetermined flight plans. Sensor positioning is supported with differential code GPS providing 1 m accuracies, and differential carrier phase positioning providing 10 cm accuracies. Either code or carrier phase measurements are employed depending on the particular mission requirements.

Through both laboratory and field experiments it has been documented that the passive microwave radiometers, and the processing / retrieval algorithms developed at the Russian Academy of Science are able to determine the following soil, water and vegetation related environmental parameters and conditions:

� Subsurface soil moisture for agricultural applications;

� Depth to a shallow water table;

Table 3 lists the operating range and errors for each parameter that can be established with the system.

_______________________________________________________________________

_______________________________________________________________________

- operating range 0.02 - 0.5 g/cc

- thickness of top soil layer

no prior knowledge is required 20 - 30 cm

prior knowledge is required 100 cm

- maximum absolute error

vegetation biomass is less than 2kg/m2 0.05 g/cc

vegetation biomass is greater than 2kg/m2 0.07 g/cc

- operating range

humid, swampy areas 0.2 - 2 m

dry arid areas, deserts 0.2 - 5 m

- maximum absolute error 0.3 - 0.6 m

- operating range 0 - 3 kg/m2

- maximum absolute error 0.2 kg/m2

- operating range 1 - 300 ppt

- maximum absolute error 1 - 5 ppt

- relative error 0.2 - 1 ppt

_______________________________________________________________________

The Mississippi River Project

The scope of work for this project was to collect and process digital passive microwave data over several sites along the Mississippi River. The areas included:

Priority Area / Description

The data was collected to establish a baseline of the spatial variation of subsurface soil moisture along the river levees. The overall purpose of the project was to: 1) conduct the survey to measure soil moisture; 2) to document the results showing the ability of the system to detect subsurface moisture variations, and; 3) to assess the feasibility of expanding this technology throughout the river system to locate and quantify levee saturation or levee under seepage during river flooding.

The microwave radiometric sensors used for this project consisted of three single-beam radiometers operated at the wavelengths of 6 cm, 18 cm, and 21 cm and a dual-channel scanning radiometer operating at the wavelengths of 2 and 5.5 cm.

The single beam radiometers are portable, stable, sensitive instruments that can be packed into a briefcase. Characteristic antenna size is 20x20x3 cm for the 6-cm radiometer and 60x60x3 cm for the 18 and 21-cm radiometers.

On the project EarthData utilized one of their Piper Navajo Chieftain aircraft to support the mission. This aircraft is a twin engine, with full mission support electronics. The radiometers were all mounted in the cabin, with the non scanning antennas mounted on a support structure designed specifically for this project.

Portable IBM compatible computers were used to operate the radiometers, and for data collection. The crew consisted of a system operator, navigator and pilot.

To adhere to FAA regulations the flights were conducted at the lowest possible altitude of 1200 feet above mean terrain (AMT). It is possible to fly lower, following special waivers from the FAA. At this altitude the spatial resolution provided by the radiometers was about 780 feet. The data capture rate was set to 10 measurements per second in each of the radiometric channels. This rather frequent sampling (compared to the characteristic frequency of signal variations caused by the soil moisture changes) permitted an artificial increase in the spatial resolution along a flightline up to 2 times.

Post flight data processing consisted of the following procedures:

(a) data georeferencing;

(b) data retrieval from the microwave radiometric values to the thematic values for map construction;

(c) map construction.

The Omnistar DGPS, integrated with an Ashtech Z-12 GPS receiver, was used to capture sensor positions during flight. This system yields positions in WGS 84, at an accuracy level of 1 meter, when used in combination with a low noise C/A code receiver such as the Ashtech Z-12. Datum transformations were performed using the USACE software package CORPSCON. The resulting GPS data file was time matched to the radiometer data file to obtain positions of the radiometer data set.

Data retrieval, or calibration, consists of two components. First, each channel is calibrated by measuring the levels of radiation from a water body and an artificial black body (measured on the ground before take-off). These objects have a known radiation value, and calibrate the upper and lower bounds of the system. Second a dual channel (6 cm and 18 cm and/or 6 cm and 21 cm) procedure is applied to determine the estimates of soil moisture and a vegetation index. This index is proportional to the vegetation biomass, based on the general type of vegetation in the region (narrow-leaf, broad-leaf crop, grasses, trees). This procedure is based on an application of the spectral peculiarities of the soil-canopy system described earlier.

The final step was to overlay the thematic radiometer data and the USACE vector data from the River Engineering and Environmental Geographic Information System (REEGIS) database. Map sheets were developed of each area and written to hard copy and CD ROM for delivery. A hard copy example is attached to this report. The brightest areas are those with the least amount of soil moisture, and the darkest blue, the highest level of moisture. Several areas can be seen that indicate moisture immediately behind (land ward) the levee. During flight EarthData also captured video imagery. This imagery was used to frame grab areas of interest and to compare the visible spectrum to the radiometer data. In the areas of high moisture behind the levee, no evidence of water could be found from the video images.

During operations the scanning system exhibited a failure on the 2.5 cm channel, preventing further data collection. The 5.5 cm channel operated throughout the mission. It was not possible, however, to achieve a satisfactory geo-reference solution for this system, therefore no data was presented. EarthData and Geoinformatic are continuing to resolve this issue for future use and application of the technology.

The single beam (non scanning) sensors worked well during the mission, and the geo-referencing allowed the presentation of a quality finished product. Minor differences were seen between the data collected at the 18-cm and 21-cm wavelengths. This is a result of similarity of the channels, both in L-band. The primary reason to use these two channels simultaneously is to protect against capture failure from interference with transmitting microwave systems. On this project there was no signal interference from electromagnetic noise.

There is a very distinct difference between the 6 cm (C-band) data and the 18 cm (L-band) data. This is a result of the sensitivity between the indices of absorption caused by the vegetation at these wavelengths, and the indices of sensitivity to the depth of water table which manifests itself via the peculiarities of sensing the capillary fringe above the level of underground water. The difference between these two imageries is a source of information about the estimate of soil moisture, depth to water table and, vegetation characteristics on this project. It is believed that the data represents soil moisture at a depth of about 5 to 10 cm for the C-band and up to 1 meter depth for the L-band data. This is based on a general understanding of the soil type, vegetative biomass, and previous experiences of the scientist involved. One important limitation is that standing water on the ground masks soil moisture below the surface. Therefore, soil moisture detection after a heavy rain would need to be delayed until after any standing water had run off, evaporated, or percolated into the soil. Rains did occur just prior to, and during, this testing resulting in data collection delays.

Because these data were collected and processed without specific on-site information on the soil properties or vegetative cover, quantitative values of subsurface moisture could not be produced. Instead the information gathered represents the spatial change of moisture over a region. It appears that about 5 gradations of moisture change were successfully depicted.

The data appear extremely valuable in the delineation of subsurface soil moisture over regional areas. No evidence was found of the ability to isolate moisture changes at a "pixel" level similar to the image interpretation of common aerial imagery. This is based on two primary factors: 1) subsurface moisture is characteristically a long wavelength phenomenon, and 2) the single beam instrument has a footprint, or ground sample distance, of 65% of the flying height. The ability to examine moisture changes at a finer resolution may be possible through a resolution of the scanning sensor geo-referencing and by flying at lower altitudes with FAA waiver.

Based on the results obtained from this project, it is reasonable to expect that the data presented could be used to delineate specific areas that require on the ground engineering examination to prevent levee under seepage or saturation during a high water event. This could be facilitated through the implementation of a systematic approach to the incorporation of this new imagery.

Data capture, hence mission planning, for the microwave technology is quite different from conventional photogrammetry. The microwave system is not directly affected by vegetative canopy, as is a conventional image, cloud cover, sun shadows, etc. It is affected by heavy rains, and should not be operated within a few days of sustained rains. In the same since, it is affected by standing water; therefore, this technique should be employed before seepage begins to pond on the landward side of a levee. It is recommended that planning for future missions coincide with the river stages and geographic areas of interest.

During this project passive microwave data was successfully collected on 6 sites along the Mississippi River. The data was presented in raster form showing 5 gradations of subsurface soil moisture without apriori knowledge of the area, or any on site ground reference. The two sets of images represent soil moisture spatial change at a depth of about 5 - 10 cm for the C-band data and up to 1 meter for the L-band data. The sensor system operates from a small aircraft making it suitable for operation throughout the Mississippi River basin.

1. Shutko, A.M. (1982) Microwave radiometry of lands under natural and artificial

moistening, IEEE Transactions on Geoscience and Remote Sensing, GE-20, 18-26.

2. Shutko, A.M. (1986) Microwave Radiometry of Water Surface and Grounds, Nauka/Science Publ. House, Moscow (In Russian). English translation is available - 1989.

3. Shutko, A.M. (1987) Remote sensing of waters and land via microwave radiometry (The principles of method, problems feasible for solving, economic use). Proc. Study Week on Remote Sensing and Its Impact on Developing Countries, Pontifical Academy of Sciences, Vatican City, 413-441.

4. Shutko, A.M. (1992) Soil-vegetation characteristics at microwave wavelengths, Chapter 5, TERRA-1: Understanding the Terrestrial Environment. The Role of the Earth Observations from Space, Taylor & Francis, London-Washington D.C., 53-66.

5. Shutko, A.M., Haldin, A.A., Novichikhin, E.P., Milshin, A.A., Golovachev, S.P., Grankov, A.G., Mishanin, V.G., Jackson, T.J., Logan, B.J., Tilley, G.B., Ramsey III, E.W., Pirchner, H. (1995) Microwave radiometers and their application in field and aircraft campaigns for remote sensing of land and water surfaces, Proc. IGARSS '95, 734-735.

6. Shutko, A., Haldin, A., Novichikhin, E., Yazerian, G., Chukhray, G., Vorobeichik, E., Agura, V., Kalashnik S., Sarkisjants, V., Sklonnaja, N., Logan, B., Ramsey III, E. (1997) Application of microwave radiometers for wetlands and estuaries monitoring, Proc. 4th International Conference on A Remote Sensing for Marine and Coastal Environment, vol.I, 553-561.

7. Shutko, A.M. (1997) Remote sensing of soil moisture and moisture related parameters by means of microwave radiometry: instruments, data, and examples of application in hydrology, NATO ASI Series, Series I:Global Environmental Change, vol.46, Land Surface Processes in Hydrology, Trials and Tribulations of Modeling and Measuring, Ed. by S.Sorooshian, H.V.Gupta, J.C.Rodda, Springer, 263-273.

Stephen R. DeLoach, P.E., L.S.

President, EarthData Technologies, LLC

18227 Airpark Drive

Hagerstown, Maryland 21742

Tel. 301.733.1176

Fax 301.733.4906

Stephen W. Ellis, P.E.

Civil Engineer

Engineering Division

Mississippi Valley Division / Mississippi River Commission

US Army Corps of Engineers

P.O. Box 80

Vicksburg, Mississippi 39181

Tel. 601.634.5910

Fax 601.634.7880