northeastern United States and eastern provinces
to the analysis of this data field, because we found
of Canada. We found in preliminary analysis that
the logarithms of temperature differences along
lesser temperature differences occurred on win-
the river to be normally distributed under lapse,
ter mornings at river level than on a nearby moun-
and nearly so under inversion. Hatch and Choate
tain, and we seek the physical mechanisms
(1929) showed that the slope of the central por-
responsible for this. A systematic approach to tem-
tion of the log normal distribution (i.e., the ratio
perature observation was developed, allowing the
of the variable at the 84th percentile to that at the
temperature to be repetitively observed adjacent
median) was numerically equal to the GSD of the
to landmarks of known elevation. We have pro-
variable. We propose that the GSD of tempera-
posed that the vertical temperature structure of
ture difference is a sensitive indicator of local
the air within the valley can be extracted from
differences in air exchange, where surface air tem-
temperature observations at two elevations. We
peratures are lower than the isentropic extension
cannot independently verify this with free air ob-
of tropospheric temperatures on a majority of win-
servation, but the inferred lapse and inversion
ter mornings. The frequency analysis of tempera-
structure is consistent with other observations.
ture differences may provide an economical
We have shown that an early morning air tem-
approach to determining exchange and disper-
perature survey can be quite quickly accomplished
sion characteristics in an area of sparse meteoro-
in a 10- 30-km area of 330-m local relief. Analy-
logical record, where activities or operations may
sis of the observed temperatures in the plane co-
require meteorological support. This proposal will
incident with the bank of the Connecticut River
be tested in Parts II and III.
provide a temperature reference plane, which dif-
fers by less than 1.2C over distances of 8 km (and
less than 1.6C over distances of 29 km) on most
CONCLUSIONS OF PART I
winter mornings. This measured temperature uni-
formity indicates that the radiative properties of
The logarithms of the temperature differences
the surface along the river and the atmosphere
observed with respect to a reference point tem-
above it are relatively uniform over this distance.
perature were normally distributed along 30 km
Magono and his coworkers used a large array
of the Connecticut River Valley. The median and
of maximum-minimum thermometers and sev-
geometric standard deviation of the logarithms of
eral tethersonde surveys to examine the tempera-
temperature differences can be calculated and
ture distribution and vertical temperature
considered in evaluating the river plane as a ref-
structure in the Moshiri Basin. We have used a
erence temperature field. The median value of
moving probe, as Benson and Bowling (1973) did,
temperature difference increased in an orderly
to survey temperature difference over a larger
way as distance from the reference point increased.
area. Our results are essentially in agreement. This
This difference increased in magnitude as snow
indicates that the moving probe technique of ex-
amining temperature structure over hilly terrain,
We conclude that atmospheric influences on
proposed in Figure 4, was apparently capable of
outgoing radiation were relatively uniform over
producing the same results as a large field of
this 30-km segment of the Connecticut River dur-
thermometers and a sonde. We used this appar-
ing this experiment period, allowing air tempera-
ent lapse or inversion structure, which was inde-
tures to achieve this uniformity. This analysis has
pendent of observations in the river basin, to
defined the resolution of a moving probe air tem-
stratify our analysis of temperature variation along
perature measuring system and the homogeneity
the river.
of surface air temperatures along the Connecticut
Several techniques of frequency analysis were
River plane sufficiently to allow the river plane
used in determining the dimensions of a region of
temperature to be used as a reference for analysis
relatively uniform air temperature above the Con-
of temperature-modifying processes in differing
necticut River. These techniques were applicable
terrain along the river.
12