This work is supported by NSF grant OPP-0085176 from the Office of Polar Programs.
STAR-Light will enable hydrologists to extend plot models of land-atmosphere energy and moisture transport processes to the circumpolar Arctic. Arctic energy balance experiments at sites that are both representative and accessible are used to develop Land-Surface Process (LSP) models, but extrapolating from these sites to the varying and mixed terrains of the circumpolar Arctic will require model calibration and validation that can be achieved only with frequent observations over broad regions beginning with spring thawing and ending with fall freezing. Once calibrated, these regional models will then serve either as improved lower boundaries of atmospheric models or as more reliable elements of integrated regional hydrology models.
At the heart of the LSP model are estimates of moisture stored in soil, vegetation, and snow. The quality of these estimates and the skill of model predictions can be significantly improved by assimilating near-daily observations of the moisture in the upper few centimeters of soil. The remote sensing hydrology community has converged upon 1.4 GHz brightness as the most effective observation for this purpose. Recent breakthroughs in radiometer technology, in LSP/Radiobrightness models, and in efficient schemes for assimilating radiobrightness have placed us on a path toward reliable long-term monitoring of changes in the amount, state, and spatial distribution of moisture stored within tundra throughout the Arctic. Essential elements of this vision are data from satellite radiometers and calibrated LSP/R models for arctic terrains.
Both the European Space Agency and NASA are developing 1.4 GHz synthetic aperture radiometers for low Earth orbit. Hydrologists are preparing for the advent of data from these instruments with extensive field campaigns to develop and calibrate LSP/R models, and to validate schemes for assimilating satellite data. The focus of these efforts has been prairie terrains. There are no proven LSP/R models for arctic terrains even though one could reasonably argue that remote sensing technologies are more vital to Earth system science in the Arctic. Mature LSP/R models for arctic terrains require collaborative campaigns involving arctic soil and snow hydrologists and remote sensing hydrologists supported by near-daily regional data from an airborne 1.4 GHz imaging radiometer. Only three such instruments are planned for the next decade - two in Europe and an enhanced version of NASA's Electronically Scanned Thinned Array Radiometer (ESTAR) which flies on the NASA P-3 - a large, 4-engine turboprop aircraft. The high operating costs of the P-3, the conflicting schedules of instruments on the P-3, and the demand for ESTAR data in NASA's many large field campaigns will greatly limit its use in seasonal and inter-annual investigations in the Arctic.
The PI's research group has developed the first example of a compact, 1.4 GHz, Direct Sampling Digital Radiometer (DSDR). The proposed STAR-Light instrument will use seven 1.4 GHz DSDR receivers configured as a 2-dimensional synthetic aperture radiometer. STAR-Light will be sufficiently compact and robust to operate in the Arctic on a light aircraft or on an Uninhabited Aerial Vehicle (UAV). A design goal is that it fit within the performance and configuration limitations of a Super Cub. STAR-Light will be designed, fabricated, and tested over a 3-year period by the Space Physics Research Laboratory (SPRL) at the University of Michigan.
STAR-Light uses the microwave remote sensing technique proven to provide the most sensitivity to soil moisture and combines it with recently developed digital technologies to produce a small, lightweight, airborne instrument capable of long term study of hydrology in difficult-to-access areas. Our objective is to make it fit on a very small and reliable aircraft such as an Aviat Husky (shown below), so that daily surveys are possible. Here's how we do it:
Microwave Radiometry at the radio astronomy frequency of 1413 MHz is the most sensitive remote sensing technology to soil moisture and ocean salinity, while maintaining the least sensitivity to interfering signals like that due to vegetation cover. Unfortunately, at this frequency the wavelength is quite large, at 21cm, and therefore decent angular, and therefore spatial, resolution requires a large aperture. For 100 m resolution from an aircraft system flying at a reasonably safe altitude of 250 m AGL, the system must achieve at most a 22 degree beamwidth. This requires an antenna aperture of about 0.5 m. While this size antenna is quite achievable in a small aircraft, to mechanically or electronically scan it would difficult. To extend the technology into higher altitudes, such as space, for large swath coverage, mechanical scanning of this technology is not possible.
Recent developments in Synthetic Thinned Array Radiometry (STAR) have demonstrated the appropriate technology path for soil moisture monitoring to a space-borne system. Indeed, the Microwave Geophysics Group was a participant in the proposal for HYDROSTAR, and the European Space Agency is currently in Phase A of the SMOS satellite.
The STAR technique creates images not by scanning, but by synthesis somewhat akin to SAR or phased arrays. By correlating the signals received by many identical radiometer receivers each with a very small antenna (and therefore wide beamwidth), the interference pattern of the imaged scene can be measured. The in-phase and quadrature component of these correlations are called visibilities, and can be converted to an image of the soil moisture scene with a Fourier transform. The most efficient layout of these identical receivers and small antennas is in the shape of a "Y", as shown below in the photo mosaic of the STAR-Light layout.
Many small identical radiometers is an ideal technology to employ on a small aircraft for long term study in a remote environment like the Arctic. A problem, however, exists in the operation of correlating the signals from these radiometers. For a system of N receivers, N^2 correlations must occur. Done in analog circuitry, this is very complex, especially considering that microwave radiometers require very precise environmental (especially thermal) control.
Digital radiometry allows the STAR technique to be used without analog correlators. The signal from each radiometer receiver is mixed down in frequency by a Local Oscillator common to all receivers, and then digitized. The resulting bit streams are then cross-correlated in digital circuits, which do not have nearly as tight an environmental requirement as does analog circuitry. This is the technique proposed for SMOS.
Direct Sampling Digital Radiometry (DSDR) is a technique developed by the University of Michigan which further simplifies the analog portion of a STAR instrument. Recent advances in Analog to Digital Converters (ADC) have pushed the upper frequency limit of the analog signal past 1413MHz, the operating frequency of an L-band soil moisture radiometer. Thus, it is possible to eliminate the distributed local oscillator circuit from the digital radiometer. Mark Fischman built the first DSDR and demonstrated its operating characteristics for his PhD at U of M. His DSDR hardware now constitutes one of the Microwave Geophysics Group's operational radiometers.
Digital technology allows further simplifications of the analog hardware. In addition to STAR-Light being the first STAR using Direct Sampling, STAR-Light is also the first STAR employing digital filtering for band definition. The STAR technique is predicated on an assumption that the different receivers used all have very similar bandpass characteristics, which, at 1413 MHz, traditionally required large, bulky, and thermally sensitive cavity filters. Digital radiometry allows the elimination of this problematic analog component. Analog filters are still required for interference rejection, but the operational requirements for interference rejection are much more relaxed than they are for band definition.
Also, careful use of digital radiometry allows the use of the Hilbert transform to generate the in-phase and quadrature signals needed for generating the visibilities. This digital technique cuts the required number of analog to digital converters in half, to one per radiometer receiver, as well as eliminating significant analog circuitry.
Thus, recent ADC technological advances allow significant reductions in the need for sensitive analog circuitry to perform Synthetic Thinned Array Radiometry. This simplification of analog circuitry, in turn, enables small, lightweight, and power efficient STAR systems to be built. Small STAR systems, in turn, enable an entirely new approach to hydrology in remote environments like the Arctic.
The STAR-Light presentation at the Microwave Specialists Meeting in Boulder, CO during November 2001.
The STAR-Light poster at ARCSS in Seattle, WA during February 2002.
The PowerPoint presentations of our design team at our Preliminary Design Review are available for download below:
If you ever need to contact me for any reason, the best by far is email.
deroo@eecs.umich.edu