During a visit with Champ Tanner at the University of Wisconsin in 1989 or 1990 we were discussing the challenge of measuring fluxes of pesticides and other trace gases. He mentioned an intriguing idea he had heard about in a paper presented at the Ag & Forest Meteorology Meetings by Businger & Oncley. They had reanalyzed some eddy covariance data sets on latent and sensible and noted that if data were sorted by the sign of vertical velocity, then the scalar fluxes were proportional to the difference in mean concentration between the positive and negative bins, multiplied by sw, the standard deviation of the vertical velocity. Champ suggested that this might be a way to measure fluxes of species for which no fast-response sensors existed. I began investigating this idea and developed a system to test the concept on CO2, with the assistance of John Norman and Bill Bland. We found that the system worked and developed the fundamental theory behind it (Baker, Norman & Bland, 1992). Some years later, Doug Cobos used the method in his dissertation research to measure mercury fluxes (Cobos, Baker & Nater, 2002).
An interest in improving the accuracy of surface temperature measurement led us to work on a new approach to infrared thermometry (Baker, Norman & Kano, 2002), in which we coupled a small IRT can to a servo. This allowed us to constantly rotate the IRT between two fields of view - one the surface of interest and the other a blackbody coupled to a Peltier block. A datalogger controls the polarity and magnitude of voltage provided to the Peltier block in order to maintain the temperature of the blackbody such that the output of the IRT is the same when its viewing the blackbody as it is when viewing the target. Under these conditions, the radiometric temperature of the blackbody and the surface are the same, and the blackbody temperature can be measured with a well-calibrated thermocouple or PRT. This resulted in a highly accurate infrared thermometer. It was deployed for several years on the WLEF radio tower in northern Wisconsin, but has not found application elsewhere, probably due to the need to control it with a dedicated datalogger.
In 2003, Tim Griffis and I began measuring CO2 and H2O fluxes continuously in two fields at the University of Minnesota's Rosemount Research and Outreach Center. This project is ongoing, and has been associated over the entire period with the Ameriflux program, a network of flux measurement sites that now extends through North America, providing publically available data to modelers and researchers around the world. Our control field, G21, has been in a conventional corn/soybean rotation throughout the period. Initially our second field was also in corn/soybean rotation, but with additional conservation practices hypothesized to favor carbon gain, e.g. - reduced tillage and winter cover crops. Unfortunately, we lost that field in 2011 when the University of Minnesota sold the mineral rights to the land and allowed a gravel company to begin strip mining it. We have since moved our second tower to a nearby restored prairie administered by the Department of Natural Resources. Our Rosemount fields have provided a rich test bed for a wide variety of research projects associated with CO2 and H2O cycling, and more recently with N2O emission. These efforts are described in more detail on our biometeorology web page, and in a series of publications.