[1] In a recent paper, Beig and Fadnavis [2001] present a re-analysis of rocket measurements of atmospheric temperature between 20 and 70 km made at Thumba, with a view to deriving a long-term trend. In their Figure 3, they compare the trend determined from these measurements with trends obtained from various other observations. Amongst the various trends compared they quote a value of 3.5 K per decade at a height of 92 km as coming from a paper published by us [Clemesha et al., 1992]. As explained below, although our observations can be interpreted as the result of a negative long-term trend in atmospheric temperature, they certainly cannot be said to indicate a specific trend at a given height. Furthermore, the extended data set which we presented in Clemesha et al. [1997], if interpreted in terms of a long-term temperature trend, indicates a rather smaller trend than the shorter data set studied in our earlier paper. In this respect it should be noted that Beig and Fadnavis include only our 1997 paper in their list of references, although the trend they quote is derived from the 1992 publication. [2] Clemesha et al. [1992] presented data on the centroid height of the atmospheric sodium layer measured by lidar. The measurements in question covered the period 1972 to 1987. A linear trend adjusted to the mean monthly centroid heights showed a decrease in this height by 49 ± 12 m yr . In our 1992 paper we commented on the possibility that the centroid height trend could be the result of a simple sinking of the isodensity levels caused by global cooling of the upper atmosphere. We further commented that a constant 5 K decrease in temperature from 50 km up would be sufficient to produce the 700 m fall in height observed between 1972 and 1987. It appears that Beig and Fadnavis [2001] have taken this value, converted it to 3.5 K/decade, and ascribed it to a height of 92 km, close to our mean centroid height, without further explanation. The purpose of this note is to point out that the trend quoted by Beig and Fadnavis on the basis of our data is unintentionally erroneous, and to clarify the possible interpretation of our data in terms of a temperature trend. [3] In a later paper [Clemesha et al., 1997] we analyzed a slightly longer data set and included a solar cycle in the trend analysis. In this paper we found a value of 37 ± 9 m yr 1 for the linear trend between 1972 and 1994, and this value should be taken as the basis of any attempt to interpret the height change in terms of a temperature trend. As a basis for discussion, we have calculated the effect on constant density levels of the temperature trend derived from the Thumba rocket measurements by Beig and Fadnavis [2001]. This is shown in Figure 1, together with the effect of a constant temperature decrease of 1.5 K from 20 km up. In both cases we have assumed zero trend below 20 km and, in the case of the Thumba data, a constant trend of 5.5 K/ decade above 70 km, the maximum height for the rocket data. [4] From Figure 1 it can be seen that the Thumba profile would result in a sinking of the constant density levels by about 730 m/decade at 92 km, and that a height independent trend of 1.5 K/decade from 20 km up would produce a trend of about 375 m/decade, close to our observed trend in the sodium layer centroid height. Thus we can conclude that, if we interpret the centroid trend in terms of a simple sinking of the atmospheric constant density levels, the observed height trend would correspond to about 50% of that derived from the rocket measurements, or to a heightindependent trend (above 20 km) of close to 1.5 K/ decade. [5] The interpretation of the centroid trend as the result of a simple sinking of constant density levels is, of course, an over-simplification. Any change in the temperature structure
Read full abstract
Feb 5, 2024
Yuchang Xun + 8
Open Access
