For some, twenty-five years of data doesn’t sound like much for use in establishing long-term trends in global temperature. But, the temperature data collected by NASA satellites corresponds with the time when the potential human impact on climate should have been greatest. It has been collected consistently from samples of the atmosphere over the entire globe. This makes it of great importance. Because the data nearly “cover the earth,” these measurements are unique among existing climate datasets. This also makes them an invaluable tool in assessing the impact of human activity on the global atmosphere. The satellite data now spans a quarter century. Happy anniversary!
The U.S. first began collecting and recording atmospheric temperature data from space platforms on November 16, 1978, a couple of weeks after the successful launch of the TIROS-N/NOAA satellite carried aloft several instruments designed to collect environmental data for the earth. One of these instruments — the Microwave Sounding Unit (MSU) — was designed to measure the temperature of the earth’s atmosphere by recording the microwave energy emitted by oxygen molecules. Such a measurement is an indication of the temperature of the molecule. These measurements have been made continuously since then using similar instruments subject to various technological improvements and launched aboard various satellites. In each instance, the new instrument became operational before the demise of its predecessor. This allowed for calibration to assure continuity of data.
Temperature observations made from space have one disadvantage; they measure only the average temperature within broad atmospheric layers. For example, a single satellite temperature measurement is made for the atmospheric layer that extends from near earth’s surface up to an altitude of 30,000 feet. The result is a measurement of the lower troposphere. But, MSUs aboard the satellites have a distinct advantage; they are able to take measurements of temperatures from over 95 percent of earth’s surface. They are “blind” only to small regions around the North and South Poles, and in places where mountains protrude through the lower atmosphere (the Himalayas).
The first MSU history of temperature measurements was published in Science in 1990 by Roy Spencer, a microwave expert at the National Atmospheric and Space Administration, and John Christy, a climatologist at the University of Alabama-Huntsville. Their analysis introduced the climatological community to the idea that satellite temperature measurements could be used to monitor global temperature change in the earth’s atmosphere. In the thirteen years since their first publication, the satellite temperature record has grown to be a centerpiece in the global warming debate. This is because it represents the most consistent, widespread, and accurate measurement of global temperature in the lower atmosphere — a region computer climate models predict should warm at the greatest rate due to the build-up of greenhouse gases. Many of us contend satellite measurements provide one the best tools available for assessing the accuracy of global climate models. As a consequence, the data set is controversial.
Spencer’s and Christy’s satellite record, with the collection and analysis of the data for November 2003, celebrates its twenty-fifth anniversary. What we’ve learned in that time is that the MSU data show global-averaged temperature in the lower troposphere to have warmed by about 0.19ºC (0.34ºF). Much of that warming has come since the El Nino of 1998 and is confined to latitudes north of 30ºN. There appears to have been little to no warming in the tropics and southern latitudes.
Figure 1 and Figure 2 are extracted from a recent report by Spencer and Christy. They depict both the 25-year-long global temperature history from the satellites (Figure 1) and the spatial pattern of the 25-year observed temperature trends (Figure 2).
Figure 1. Monthly satellite temperature anomalies from November 1978 through November 2003 (from Spencer and Christy, http://www.uah.edu/News/climate/25years.pdf).
Figure 2. Pattern of temperature change as observed by satellites for the period November 1978 through October 2003 (from Spencer and Christy, http://www.uah.edu/News/climate/25years.pdf).
The satellite data record indicates the atmosphere has warmed up over the last twenty-five years (1) less than the amount calculated using observations taken at the earth’s surface, and (2) much less than the amount calculated by most global climate models, which incorporate observed changes to the atmospheric system (i.e., volcanic eruptions, solar variations, greenhouse gas increases, aerosol changes, etc.). The discrepancy in these measurements casts doubt on what scientists really know about what is going on with regard to “global warming.” It raises three immediate questions: are the climate models functioning correctly, is the satellite data accurate, and/or is there a flaw in the surface observations? As a consequence, Spencer’s and Christy’s analysis of the satellite data is given close scrutiny by several research teams who suggest there are adverse influences on the satellite observations caused by such things as orbital decay, orbital drift, and intersatellite calibration. They claim Christy and Spencer do not properly account for such effects.
On the contrary, Spencer and Christy have carefully addressed each expressed concern and done so in great detail. They have also demonstrated close correspondence between their calculated temperatures and those independently observed using radiosondes to transmit data from the weather balloons that are launched twice daily around the world and used to gather meteorological information for daily weather forecasts. Similar close correspondence has not been demonstrated for alternative representations of Christy’s and Spencer’s satellite data. Therefore, their data remain the de facto standard to beat.
We conclude that the satellite data is all but completely verified and the discrepancy between datasets suggests problems with the surface data, the computer models, or a bit of each. Surface temperature measurements will respond to influences that are not related to large-scale climate changes: things such as urbanization, landscape change, changes in agricultural practices, movement of instrument location, and changing times of observation. There have been many attempts to account for such influences, but most are not easily quantifiable and therefore are difficult to adequately account for. Furthermore, many regions of the earth are sparsely sampled. In regions like the oceans (which comprise most of earth’s surface), observations are extrapolated from a single observation point and are assigned to represent large geographic regions. This less than ideal situation illustrates an inherent and not easily overcome flaw in the surface temperature records.
We also note that in the regions where surface temperature measurements are most plentiful, land areas in the United States, Europe, Russia, China, and Australia, there appears to be greater agreement between the surface measurements and the satellite measurements. This correspondence weakens where measurements are sparser, suggesting that, in calculating large-scale temperature trends (hemispheric or global), the satellite data may better represent real conditions than surface data.
The computer models introduce another source of error. On one hand, if the surface data is proven to be correct, the climate models would indicate the lower atmosphere should warm as fast as, or faster than, the surface. Yet surface temperatures have been observed to be warming faster than temperatures in the lower atmosphere. This fact alone indicates a major problem in the models. Yet, on the other hand, should the surface data prove to contain non-climatic warming elements which, when removed, bring the models more in line with the satellite measurements, only then could one claim the climate models better represent the vertical structure of the warming, but in so doing would be greatly over-predicting the rate of observed warming by about a third or a half. Whichever the case may be, such inaccuracies cast doubt on the veracity and subsequent utility of climate models as a tool for assessing future climate. If the models cannot accurately capture the known behavior of the earth’s climate, they simply cannot be relied upon to make accurate future projections.
In the grand scheme of things, twenty-five years of data represents a tiny fraction of earth’s climatic history. But to anyone who has been involved in the greenhouse debate, those twenty-five years of satellite temperature observations provide a wealth of information. So, in this season of celebration, let’s add a toast for Spencer and Christy’s tireless efforts and to another quarter-century of satellite data collection as fruitful as the last.
References:
Spencer, R.W., and J.R. Christy, 1990. Precise monitoring of global temperature trends from satellites. Science, 247, 1558-1562.
Spencer, R.W., And J.R. Christy, 2003. Global temperature Report, 1978-2003. http://www.uah.edu/News/climate/25years.pdf
