U.S. DEPARTMENT OF COMMERCENATIONAL OCEANIC AND ATMOSPHERIC ADMINISTRATION

NATIONAL WEATHER SERVICENATI.ONAL METEOROLOGICAL CENTER

OFFICE NOTE 238

Statistical Survey of Wind Sensor Colocations

Charles VJcekDevelopment Division

JUNE 1981

This is an unreviewed manuscript,. primarilyintended for informal exchange of informationamong NMC staff members.

"What is truth?" - Pontius Pilate (John 38:18)

1.0 INTRODUCTION

This report describes the statistical analyses of various wind

sensor colocations drawn from radiosonde, aircraft, ASDAR, and satellite

data collected sinde October 1977. Two quantities were primarily used

for analysis: the monthly mean wind speed difference (BIAS) between two

sensors in a particular type of colocation, and the monthly root mean

square vector difference (RMSVE) between these two sensors. Further

analyses were made of the BIAS and RMSVE as a function of separation

distance (or difference in observation time) between the colocated sen-

sors, altitude of the colocations, and time. A rough comparison among

sensors is given, along with a greater in-depth study of colocations

involving Japanese satellite data.

For a number of reasons which are given in this report, it might be

considered both futile and foolish to make any bold statements about the

quality of various sensors, either in an absolute sense or relative to

each other. If one sensor were "perfect" (as radiosondes might be con-

sidered), it dould be used to calibrate the others. As it is, an attempt

at meaningful comparison is apt to be drowned in a sea of caveats.

Despite these problems, several techniques were used to provide clues

that might aid in evaluating these sensors.

2.0 HISTORY AND METHODOLOGY OF DATA COLLECTION

Collection of colocated data began in October 1977 when aircraft

and radiosonde reports were gathered and paired after testing for a

proximity threshold of 1 hour observation time and within 3 latitude

degrees ( 333 km) of each other. ASDAR-RAOB colocations were tabulated

!

2

separately as ASDAR data became available. In September 1978, satellite

colocations with RAOBS and aircraft were included There were four

satellites GOES-A GOES-B, a Japanese geostationary satellite and a

European geostationary satellite (hereafter called EURSAT). The European

satellite ceased to function in November 1979. In January 1979 the

colocation "window" was expanded to 3 hours, 3 degrees. Colocations

between observations of the same type sensor began in December 1979.

Data saved on tape c6onsists of the position and altitude (pressure

level) of one of the two colocated sensors. Only the differences in

temperature, u- and v-wind components, wind speed, and wind vector

were stored on tape. Finally, the actual horizontal and temporal separa-

tion between the two sensors were recorded, along with a numerical code

identifying the type of colocation. In October 1979 the average wind

speed of the two colodated sensors was included in the data collection.

RAOB data was interpolated (linearly with respect to log p) to the

level reported by the sensor colocated with it. Where two single-level

sensors were involved, a vertical proximity of 2000 feet or less was

required. The actual vertical separation was not recorded. The standard

atmosphere was used to adjust the temperature of one sensor to the level

of the other; however, little has been done so far in compiling temperature

statistics. Single level reports having climatologically unrealistic

temperatures for the reported location were deleted. Temperature differ-

ences between two colocated sensors exceeding 25°C were also summarily

tossed, as were wind speed differences exceeding 50 m/sec. No other

quality control techniques were used.

3

3.0 OBSTACLES TO SENSOR EVALUATION

The purpose of evaluating a sensor is to determine whether it provides

sufficiently accurate data to be useful to analysts and forecasters. A

related purpose is to isolate the cause of any problem that exists and

correct it if possible. When colocations are used as a means of evaluation,

a sensor being tested should be colocated precisely with another sensor

that is perfect, so that the BIAS and RMSVE for the test sensor is with

respect to the true wind. Since such ideal conditions do not occur, any

difference between the two sensors' wind measurements can be due to a

variety of causes (table 1) which are difficult to isolate. A simple

straight forward evaluation of such colocations would tend to be inconclusive

or misleading.

There are a few additional pitfalls inherent in the sampling. The

most obvious of these is the large variation of sample size for different

types of colocations, ranging from fewer than 30 per month for some

aircraft versus satellite colocations to more than 6,000 per month for

aircraft versus RAOB. These numbers are for the 3°, 1 hour "window"

from which most statistics given in this report are drawn. In some

instances opening the window to 3°, 3 hours would have greatly enlarged a

pitifully small sample to a statistically reliable size.

Another tricky aspect of sampling is that (for example) 500 pairs

of colocations within the 3°, 1 hour window may in fact include 5 pairs

within 1° and 1 hour--or 50, or 450. Obviously a set containing 450

pairs within 1° and 1 hour is going to produce better results than a set

containing only 5 such pairs. Fortunately, such skewed distributions

are rare. Finally, one must consider the distribution of the colocations.

4

There is little or no overlap in the regions covered by the four satel-

lites in this study. Due to the size of the "window" required to obtain

a reasonable sample, wind shear is very much a factor in this study, and

one satellite may dover a more strongly sheared region than another.

Colocations involving airkraft, RAOBs, and ASDAR may have the same problem,

though to a lesser degree.

4.0 ANALYSIS TECHNIQUES USED

Certain analysis techniques were used in attempts to alleviate some

of the problems discussed earlier. The basik strategy was to try to

isolate some of the variables that are responsible for wind speed dif-

ferences between colocated sensors, then piece the results together to

develop some kind of overall picture. This method appears to have had

some limited success. A description of these techniques follow:

4.1 Adjacent gridpoints of an operational wind analysis field were treated

as "colocated sensors" and the RMSVE calculated as a function of the separation

distance between gridpoints (Fig. 1). The purpose of this technique was

to determine the probable overall RMSVE value of a real wind field, i.e.,

the "meteorological" contribution to wind vector differences between

colocated sensors. This value appears to be roughly 5 m/sec, or 25%

of the mean wind, for colocations having a horizontal separation of 3°

(333 km) and under.

4.2 The BIAS for a sufficiently large number of colocated sensors of

the same type (i.e., GOES-A vs GOES-A) should be zero. Non-zero values

are random fluctuations the magnitude of which should be in some degree

inversely proportional to the sample size. These fluctuations were used

to construct an empirically derived set of "confidence limits" for BIAS

as a function of sample size. The resulting graph (Fig. 2) should give

5

a pretty good idea of what the significance threshold of BIAS would be

for a given set of colocations between different types of sensors. For

example suppose that an airdraft-satellite pairing has 80 colocations

and a mean speed difference of 2.58 m/se6. The graph indicates that for

80 colorations, a difference of up to 3.4 m/sec can occur by chance; therefore

a difference of only 2.58 m/sec is not significant.

4.3 Monthly tabulations of BIAS and RMSVE were obtained from "window boxes"

ranging from 1°, 1 hr to 3°, 3 hr (Fig. 3). If the wind vector differences

were only meteorological, the RMSVE should increase with increasing window

size. What kind of "window box" distribution might we expect? Disregarding

the temporal variation, consider Fig. 4, in which we have a wind field with

parallel flow and uniform speed shear across the flow (as might be found in

a small area on one side of a jet stream). Here V everywhere is zero,

du/dx=0, and du/dy=2 m/sed/deg. If every point is "colocated" with the

center, the RMSVE varies linearly with the radius of the circle (i.e.,

the maximum separation), with a value of Y*(du/dy)/2, where Y = R.

In this example RMSVE=1 m/sec for 1°, 2 m/sec for 2°, and 3 m/sec for

3° . The technique described in 4.1 and accompanying Fig. 1 suggest that

this model is valid for colocations with 3° separation at least in a

smoothed wind field. A big fly in the ointment, however, is that there

are often not enough colocations in the 1°, 1 hr. "box" to make this

technique useful.

4.4 RMSVE and BIAS statistics of colocations between different sensors

(see Figs. 5-10 for time series graphs) are the "meat and potatoes" of

this study and will readily reveal differences between sensors. However

no further conclusion may be revealed with these statistics alone, and

even the degree of apparent difference may be modified when viewed against

6

the background of other analyses.

4.5 RMSVE statistids of "same sensor" colocations is a good test of the

internal integrity of a system. See Figs. 11 and 12 for time-series

graphs. There is no real "BIAS" to bloat the RMSVE values, and a large

RMSVE cannot be blamed on "the other sensor". Unfortunately it was not

practical to test RAOBS in this manner.

4.6 "Proportional" BIAS and RMSVE values were computed since October

1979. Here the BIAS and RMSVE values were divided by the "observed"

wind speed (average of the two sensors). All other factors being equal,

wind speed error (BIAS) should increase with wind speed, and RMSVE is

increased by strong winds due to the greater signifieance of small direc-

tional differences.

Theoretically, RMS(VE/S) should be more accurate than RMSVE/S.

However, RMSVE/S was computed first, and a comparison sample of RMS(VE/S)

showed little difference, at least in the relationship between sensors.

Hence the RMSVE/S data shall be used in the discussion.

4.7 Special techniques used to evaluate 6olocations involving Japanese

satellites will be described later, during discussion of that evaluation

and its results.

5.0 RESULTS OF ANALYSES

5.1 BIAS: Figs. 5(a, b, c, d) each show a time series of BIAS values

for a satellite vs RAOB, airdraft vs satellite, and aircraft vs RAOB

(within the domain of the satellite). The sample sizes are not given

here but only the BIAS values of Japanese satellite or European satellite

(EURSAT) vs RAOB or aircraft are significant, with a few minor exceptions.

EURSAT is now defunct so only limited data is available. An in-depth

7

discussion of the Japanese satellite BIAS will be given in the next

chapter.

5.2 RMSVE of sensors vs RAOB: Fig. 6 shows a time series of RMSVE of

various sensors vs RAOB. Remember that colocations having a high BIAS

will have inflated RMSVE values. Above 500 mb ASDAR appears to have the

best agreement with RAOBS, with other aircraft having a parallel RMSVE

values about 3 m/sec higher. Japanese satellites show a February maximum

as a result of high BIAS values; if the standard deviation were taken instead

of RMS, Japanese satellites would probably have the lowest values along with

ASDAR. The remaining satellites appear to have net values between those for

aircraft and ASDAR. The limited EURSAT data available is for an earlier

time period but the mean value of about 14 m/sec is close to the GOES-A

and GOES-B values for that period.

Below 500 mb (Fig. 7) aircraft make a relatively poor showing, but

the sample size (<20) is too small to be significant. Japanese satellites

also have relatively high RMSVE, with the usual winter maximum. ASDAR,

GOES-A, and GOES-B are pretty close together.

When the RMSVE is expressed as a percentage of the mean wind, a

different pattern appears (Fig. 8). ASDAR is still best above 500 mb

but the rest are fairly dlose together although GOES-A is a little worse.

For the Japanese satellite, the winter maximum is replaced by a summer

maximum. EURSAT data of this type was available only for October and

November 1979. The values of 68 and 71 percent, respectively, were at

least 10 points higher than for any other colocation at that time.

Below 500 mb (Fig. 9) ASDAR again wins by a wide margin. Ignoring

aircraft (due to the low sample size), we see that the time-series lines

run fairly parallel with GOES-A, GOES-B, and Japanese satellites finishing

8

in order after ASDAR. This is not too much different from the non-proportional

RMSVE below 500 mb.

It should be noted briefly that RMSVE/S values above 500 mb for

satellites versus aircraft (Fig. 10) are very similar to those for satel-

lites versus RAOBs. The number of colocations per month generally ranges

from 50 to 400.

Two things should be pointed out in digesting these results. The

first is that the results simply imply the degree of agreement with RAOBS

which means little unless RAOBS are (or are assumed to be) the best wind

measuring systems available. The sedond relates to the numerical values

of RMSVE/S which appear to be rather high, especially below 500 mb. Adtually

they are fairly reasonable. Using Fig. 4 in the example given in 4.3, consider

what happens if the wind is calm at the center. This gives the lowest possible

mean absolute wind speed and the resulting RMSVE/S is 2.00 (i.e. 200%). This

is the highest possible value that can be obtained legitimately.

In a region of light winds and small scale features such as the tops

of subtropical ocean stratus decks measured by satellites, these conditions

may be approached. Because of the way S is computed, using wind measured by

both sensors, the value of S computed from Fig. 4 would be halfway between

the actual mean wind if the u component were uniformly positive or negative.

If S in the example is 6 m/sec RMSVE/S at 3° is 0.50 (50%). This is

close to the medium value for the graph of 3° colocations above 500 mb,

but there S ranges from 20 to 40 m/sec, so the mean horizontal shear

would have to be from 10 to 20 m/sec over a distance of 333 km. (3°

latitude). This does not seem to be unreasonable, and any remaining

error (attributed to the sensors) should be tolerable. Unfortunately,

the 50% value also shows up in 1° colocations; a shear of 10 to 20 m/sed

9

over only 111 km seems rather large.

5.3 RMSVE of sensors colocatedwith other sensors of the same type: This

type of colocation has the advantage of eliminating the effects of BIAS and

the question of which of the two sensors is responsible for high RMS values.

Other factors remain of course, but a direct comparison between sensors

is more meaningful than those with RAOB as a common partner.

Since only Japanese satellites have a respectable number of

colocations below 500 mb, no graph will be presented for that level.

Above 500 mb (Fig. 11) aircraft are far worse than the others, which are

nearly equal. ASDAR colocations are rather few (dying out completely

at the end), accounting for the jumpiness of RMSVE values.

For RMSVE/S (Fig. 12), the sensors are nearly equal in winter. In

summer aircraft is worse, ASDAR is better, and the three satellites are

nearly equal half way in between. ASDAR values near the end should be

taken with a large block of salt since there are fewer than 10 colocations

per month.

It is interesting to note that the RMSVE values for these colocations

are significantly lower than corresponding values for any sensor vs RAOB,

despite the fact that these are single level sensors and cannot be interpolated

to each other. Also, the RMSVE/S values for satellite vs. aircraft

colocations are higher than for aircraft vs aircraft or satellite vs

satellite. This pattern suggests the possibility of some intrinsic

differences between sensors but there may be other factors involved

which are more difficult to identify or isolate. One possibility is

that vertical separation distances are smaller for same sensor colocations.

5.4 "Window box" patterns - It was mentioned earlier that the contribution

of horizontal wind shear to the RMSVE of colocated sensors should vary directly

10

with the distance of maximum allowable separation. For example, if the RMSVE for

1° coloCations is 6 m/sec, the RMSVE for 2° colocations should be 12 m/sec, and

for 3° colocations it should be 18 m/sec. This relationship assumes the absence

of significant small scale features in the wind field.

A casual view of observed patterns shows that such a relationship

almost never occurs. In many cases, there are not enough colocations

to establish a meaningful pattern. The following discussion is therefore

confined to the most common colocations such as aircraft-RAOB, aircraft-

aircraft, and Japanese satellite vs RAOBS or other Japanese satellites.

Sample RMSVE values given in Table 2 are 6 month averages over the last

half of 1980 for 1 hr coloCations above 500 mb.

The ratio of RMSVE (3°) to RMSVE (1°) should be about 3:1. The actual

ratio ranges from 1:1 to 3:2. Same sensor dolocations have better (higher)

ratios than RAOB colocations. ASDARS are better than aircraft or satellites;

the latter two are virtually equal. The RMS values for 2° colocations are more

or less midway between the values for 1° and 3° colocations.

Small scale features in the wind field sensors may be a significant

cause of such a large difference between theoretical and a6tual ratios.

Uncertainty of actual horizontal positions of the sensors could also be responsible

for these results, as well as temporal variations (1 hour can be equivalent to

1/2 degree latitude). In the case of ASDAR vs ASDAR, positions are supposed

to be pretty precise; however, the temporal window was 3 hours. For the 1 hour

window, a casual glance at ASDAR vs ASDAR suggests a ratio that may be as good

as 2:1. Finally, the importance of vertical wind shear cannot be discounted;

its contribution to the RMSVE may range from 4 to 10 m/sec. This appears

to be a very large fraction of the values given in Table 2, and perhaps

actual vertical separations between single level sensors are significantly less

11

than for the samples described below.

5.5 A spedial "colodation" was made of the two adjacent radiosonde sounding

levels immediately above and below the level of a colocated sensor. This would

determine the Contribution of vertical wind shear to the overall RMSVE, particularly

for two single-level sensors. Since the horizontal and temporal separation

is virtually zero and instrument characteristics are not a factor, almost all

of the RMSVE is due to vertical wind shear. The results here were poor because

adjacent radiosonde levels were seldom within 2000' of each other (the limit

of vertical separation for single level sensors). Unfortunately, the

actual vertical separation was not recorded. The low number of usable reports

caused considerable jumpiness in the statistics, but the average RMSVE appears

to be about 7 m/sec, and RMSVE/S averages about 20%.

6.0 JAPANESE SATELLITES: AN IN-DEPTH STUDY

6.1 Introduction: Fig. 5c showed a very interesting pattern of BIAS

above 500 mb involving Japanese satellites, airdraft, and RAOBS. A very

large seasonal fluctuation indicated that Japanese satellite winds were

considerably slower than aircraft or RAOB winds during the northern

hemisphere winter, while aircraft vs. RAOB 6olocations in the region

covered by Japanese satellites showed little BIAS (aircraft winds tend

to be slightly faster than RAOB winds in winter).

These winter winds were quite strong, particularly in satellite-

aircraft colocations, but even the normalized value (BIAS/S) was very

high. Two explanations for this phenomenon are possible: 1) Japanese

satellite cloud top altitudes are biased (in regions of strong vertical

shear), or 2) gravity wave phenomena in the cloud tops mask the true

wind speed. The regularity, magnitude, and persistence of the BIAS

pattern seems to favor the first explanation as the primary cause.

12

Robert Hirano (unpublished report, 1981) has been working with limited

cases in applying a "level of best fit" technique, in which he found

that satellite cloud top measurements tend to average about 2 km too

high. Frederick R. Mosher described cloud height assignment techniques,

admitting that these height assignments were probably the largest source

of error in cloud-tracked winds (p. 58, Systems & Techniques for Synoptic

Wind Finding, Atmospheric Technology, Number 10, winter 1978-79, NCAR).

This chapter describes some other ways of investigating the problem.

6.2 Data: A check of actual distribution of Japanese satellite-RAOB coloca-

tions revealed that many of them are concentrated over mainland China,

particularly during the northern hemisphere winter when two-thirds of

all the colocations are located there. A listing of RAOB-Japanese

satellite BIAS for each 10° longitude-latitude "square" revealed large

positive values (satellite winds slower) in the mid-latitude southern

hemisphere winter. Finally, a time-series map of wind speed (average of

Japanese satellite and RAOB) vs BIAS for the northern hemisphere was

constructed (Fig. 13). This map covers all colocations (3°, 3 hr) at all

levels.

Japanese winds are slowest with respect to RAOBS in regions of strong

winds during the winter. They are nearly equal to radiosonde winds

during the summer, regardless of wind speed, and are often somewhat faster.

It is quite apparent that the strength of wind shear with height, rather

than absolute wind speed, is the key factor in the Japanese satellite

BIAS. A strong jet stream with strong vertical wind shear is usually

present over mainland China in winter. Mosher (ibid) indicated that the

emissivity of thin cirrus layers (such as found with a jet stream) is

considerably less than unity, causing problems with height assignments.

13

In dense cirrus, such as are found with convective and tropical systems

during the summer, emissivity is close to unity, permitting fairly accurate

height assignments.

The un6orrected emmissivity of thin cirrus, making the cloud appear

warmer than it is, would result in a dloud height assignment that is

too low. However, Hirano indicated in his report that most often these

assignments were too high. If this is true, the problem appears to be

one of over dorrection.

An attempt was made to provide a profile of BIAS vs height; this

proved unsatisfactory because most colorations were clustered in several

narrow zones with no real continuity. Another attempt was made to correlate

individual wind speed differences with actual vertical wind shear (Fig. 14).

Since actual heights were not always available, wind shear had to be

with respect to pressure. This problem was alleviated somewhat by plotting

results for a single layer 50 mb thick. Horizontal wind shear also

muddied the picture; the failure of Fig. 14 to show any correlation between

wind speed differences and vertical wind shear could easily be attributed

to these two factors. This was one of those tests where a definite

correlation (in face of the obstacles) would have been significant but

lack of a correlation would not.

6.3 Conalusion and remarks: There is evidence that cloud top height

measurement errors are largely responsible for slow Japanese satellite

winds; this evidence is strong but largely Circumstantial. A single

BIAS vs wind speed profile indicated that GOES-A and GOES-B exhibit the

same behavior as the Japanese satellite, but there are too few colocations

for an in-depth study. More uniform distribution of GOES colocations

apparently prevented any marked-BIAS pattern from showing up on the time

14

series graph. An interesting footnote is that aircraft vs RAOB colocations,

used as a control in BIAS vs height profiles, indicated that aircraft

winds tend to be up to 6 m/sec faster than radiosonde winds in the region

covered by Japanese satellites during the winter when winds were about

60 m/sec. Otherwise aircraft winds were little affected by wind speed

or season.

7.0 SUMMARY AND CONCLUSIONS

7.1 The "window box" theory and the graph shown in Fig. 1 predict that

RMSVE of colocated sensors varies directly with separation distance

between them as long as that distance is significantly smaller than the

scale of wind patterns.

7.2 Observed "window box" patterns rarely even approach theoretical patterns.

There are five possible causes:

a) For single level coloRations, vertical wind shear alone can cause an

RMS of about 7 m/sec (above 500 mb) thus putting a lower limit on observed

RMS values.

b) For satellite colocations with RAOB, apparent height misalignment

combined with vertical wind shear can produce the same results described

above.

c) Smaller sdale wind features may be present often enough to influence

the pattern. These may be more prevalent at lower altitudes.

d) Uncertainty of position (and horizontal separation) of sensors.

e) Temporal separation may have a small effect.

7.3 Observed RMSVE/S values appear to be reasonable for most 3° coloca-

tions, considering the amount of horizontal wind shear implied by actual

winds and "window box" theory.

15

7.4 No direct evidence is available for judging the quality of radiosondes

with respect to other sensors. There is some indirect evidence to indicate

that radiosondes may be no better and no worse than the others: RMSVE/S

values for aircraft vs. satellite colocations are approximately the same

as the values for colocations involving radiosondes.

7.5 In colocations above 500 mb involving radiosondes, ASDARS have the

lowest RMSVE, Japanese satellites the highest, and the rest pretty much

in between. When RMSVE/S is used, ASDAR continues to have the lower

values. Below 500 mb, ASDAR has the lowest values for either RMSVE or

RMSVE/S, followed by the three satellites. Aircraft samples are too

small to be significant.

7.6 In "same sensor" colocations above 500 mb, aircraft have significantly

higher RMSVE and the rest are about the same. RMSVE/S values of "same

sensor" colodations are probably the best indicators available of the relative

quality of various sensors. These indicators show that there is little

difference in winter while in summer ASDARS have lower values, aircraft

have higher values, and the satellites are in between. Comparison below

500 mb is not possible due to lack of data.

7.7 In the limited data available, European satellites are rather poor

with respect to the other sensors, particularly as measured by RMSVE/S.

7.8 Japanese satellite winds are much slower than aircraft or radiosonde

winds in the northern hemisphere winter, particularly when winds are

strong. In summer this BIAS is nearly independent of the wind speed and

differences are small; satellite winds may be slightly faster than radio-

sonde winds. Most Japanese satellite colocations are over mainland

China, suggesting that a combination of erroneous cloud top measurements

and strong wind shear in a jet stream is responsible for the phenomenon.and strong wind shear in a jet stream is responsible for the phenomenon.

16

Other attempts to verify this relationship have proved to be inconclusive

but support can be found in work by other investigators. Wind speed vs

BIAS profiles suggest that GOES-A and GOES-B behave in a similar manner

but is less obvious because of the greater dispersion of colocations

(fewer occuring in jet stream areas).

7.9 In a nutshell the following statements can be made about colocations

(liberally prefixed by the words probably, perhaps, and maybe):

a) A certain consistency of results suggest that ASDARS are best,

GOES-A and GOES-B are close behind, followed by aircraft. The limited

EURSAT data indidates rather poor quality. The Japanese satellite reports

did poorly because of the BIAS problem, but they would rival ASDAR if this

problem were corrected

b) There is a faint hint that radiosondes may be in the same general

category as the others sensors with respect to the quality of wind reports.

c) The ratio between the best and the worst sensors appears to be

about 3:2. This ratio would exceed 2:1 if the wind shear effects were

subtracted out. These wind shear effects appear to be about the same

magnitude as apparent differences between sensors.

d) Satellites seem to have a problem with cloud-top (wind level)

assignments. If this problem were rectified, satellites could be the

best sensors available.

e) All sensors, with the possible exception of EURSAT, appear to be

adequate. Japanese satellite data should be useable with even a rough

empirical correction for BIAS (see fig. 13).

17

TABLE 1

POSSIBLE SOURCES OF DIFFERENCES

BETWEEN WIND MEASUREMENTS BY VARIOUS

COLOCATED SENSORS

A. Real (Meteorological) Differences

1. Horizontal Shear

2. Vertical Shear

3. Temporal Variation (Intrinsic & Advective)

B. Position Errors + Meteorological Differences

1. Horizontal Position Error by One or Both Sensors

2. Vertical Position Error by One or Both Sensors

3. Temporal Error (obs time) of One or Both Sensors

C. Scaling Errors

1. Excess Resolution + Small Scale Activity (thunderstorms, etc.)

2. Insufficient Resolution of One Sensor in Jet Stream

D. Instrument errors

1. Bias by One or Both Sensors

2. Calibration Problems - One or Both Sensors

3. Intrinsic RMS in Instrument Response (Accuracy)

4. Triangulation Errors (RAOB)

E. "Transmission" Errors (time, position, or wind vectors)

18

TABLE 2

RMSVE VALUES AS A FUNCTION OF SEPARATION DISTANCE

July-Dec 1980 Average for 1 Hr Colocations Above 500 mb

Colocation

Aircraftvs

RAOB

ASDARvs

RAOB

Aircraftvs

Aircraft

ASDARvs

ASDAR

Japanesevs

Japanese

Separation

10

2°

30

10

2°0

30

1°

2°

30

10

2°

30

1°

2°

30

Sat.

Sat.

RMSVE(m/sec)

12.9

13.3

14.2

7.28.4

10.1

10.5

11.8

13.3

5.1

6.27.8

6.06.97.8

Number (per month)

1000

100-300

>10000

100 (3 hrs)

3000

r- -- 7--- l

i ;G R it I..

I . . _

!14 I R i I 01

!I-_ : -.t-l,0-:p --t

1.- , V _ $i t Th4 _

it !J 4- -.l74 I -... ._I 1 I

~}I }..ij - ii-.i 4 ~ i I 4 .~-v L .~

i!i:-:-:: .....

0:t - t11",1''4 -d

, 0t- } ~'t-

7rt;i~

!__ :~x! 1

V .\ i I __LAt ~-W.. II,- , I EP )'T. k -1. .i- 1 I J - ,^- .1 +'l l,-

i t.I' ' 1 ~ :.-. S ru

I~~ 1.- _ , 7 - -

h __-

t It

I-] I41

! 1-:-! !t--?i :!I!l

]].1tC' i ]--1-

T-t -

;I

L_ i_

F , - -

. LI I-I-v-i

-~ IIi T

r7 .. :i

-I

-I--

I-T-!I

i-- /I-rI

I I I III I

II

0 - . 7i iII , 1 2 .. 14' - i

T htMFT f t0

']35

svS

30

E

C

IS.

10

,1 hL

I I I

ITI

-1-- L-, .- IL-~

I

I

i

1!

I

I

II

/ t'7

I-t- I - L -L11:i D i- I1

rr t + - , LIt ,-

tt,�#_

__ - l - _r - -. -_ - I

wob4E~~~- Lt-

A. L ~~, I, I-z .1

I/C

! I i I

I I1T

4t-

t

Ill

I I . i

X t I

171

, 1_ ,

t 1 tl -1

III I

i'1

I y - --~1 A

91

tt I -+- -- I ' 14471 174.

S0oa 6aaoo 700o

t

- ' ......

800 9oo0 Ioc+

PAN, WI-

f. 4

-.

/iK -U1 1 5 .

14_

i .ii _

4jI 411 -IH -.

bit 1 -i: I 10

S t I t0 4 i T I. v P D - J -t TT ,:l.

(I 3S?)

1(6k

.I I -

I .]:"

II

I

/ Od 300 400

1) AR.CZLU'rz hivrI5:l

aO0:

.,** I;N(IGRAM .Q2. (C. VLCE.K, 9 I'OV. 1~79): 5urrol~r;Ax L sri:SAIISTIC FORvM 81u40000 T lOu1050OU0 ***

NUMBUER UF COLOCATIOrNS OF EACH TYPE WITHIN THE INUICATED SPACE-TIME WIJDOW LISTED EELOW

ULG HR

1

. 1

1., . 2. 2

2

2

3

3

3

n 1 2 . 3 4 a 6 7 8 9 O = AIRCRAFT $VS RO/3

11 " ~,6qA2 1405

3 ' 2254

1 2261

2 4822

3 7483

1 5448

2 1165q

3 18702

89

160

2682H9

566

924

549

1086

18 8

0

43

44

3

218

2195

533

538

20

26

26

76

116

116

159

269

269

299

300

300

1269

1287

1287

2914

2952

2952

4

10

21

25

53

92

45

111

196

2239

52

70

153203163

25

445

26

43

76

g0

143

259160

301511

o o 0 I ADAqR VS 9/0o

° ° 2 =GOES-A V3 RAOB

o o0 0 3GOES-, VS AoB

o o J,- SAT v$ RloB

0 0O 0 ~, O G;OES -A V£ AICIRC1,4FT

0 0

0 0 6 2GOES8 z A"GiRcI'r

o o 7.,'-SA7 VS AIRCRAFT

MEAN SPEEU LRROR STATISTICS FOR EACH TYPE OF COLOCATION WITHIN THE INDICATED SPACE-TIME WINDOW LISTED BELOW

DEG HR 0

1 1 0.97

1 2 1.07

1 3 0.99

2 1 0.54

2 2 0.79

2 3 0.68

3 1 0.20

3 2 0.36

33 3 0.35

1

0.82

0.83

.1.06

0.84

0.52

0.860.400.43

0.74

2 3

0.0 2.42

1.78 2.90

1.99 2.90

3.10 0.761.47 1.11

1.51 1.11

4.22 0.81

0.94 1.10

0.97 1.10

4 5 6 7 8 9

-4.70

-4.73

-4.73

-4.93

-4.88

-4.88

-4,93

-4.91

-4,91

-0.90

1.33

3.600.59

-1.35

-0.35

-0.55-1.50-0.7e

-1. 99

-1.45-1.37

-0.03

1.03

1.19

-0.44

0.48

0.81

2.38

4.70

5.314.67

6.36

6.44 -

6.76

7.22

6.58

0.00.00.00.00.00.00.0

0.00.0

0.00.00.00.00.00.00.00.0

0.0

RMSVE STATISTICS FOR EACH TYPE OF COLOCATION WITHIN THE I6DICATED SPACE-TIME WINDOW LISTED BELOW

2 3

0.0 12.97

11.85 12.7011.83 . 12.70

16.63 11.65

11.80 11.94

11.79 11.9.4

13.71 12.13

12.75 12.51

12.84 12.51

4

14.47

14.4714.4715.0114.98

14.98

15.6515.61

15.61

* 6

7.0911.50

14.5313.47

12.21

13.5512.9113.2015.1!

15.4012.7512.27

14'64

14.14

15'0715.27

14.9315.81

7 8 9

12.45

13.94

14.9515.26

16.0015.68

17.60

18.0317.07

0.00.00.00.00.00.00 .0

0.00.0

0.00 .0

0.00.00.00.00.00.00.0

FIG. 3

ITUDE ANC HR IS MAX TIME SEPARATION IN WHOLE HOUr

DEG

1

1

.1

2

2

2

3

3

3

HR

1

2

3

1.

2

.3

1

2

3

0

11.99

12.10

12.29

12.20

12.4712.60

13.3913.4913.66

1

6.47

6.46

6.75

9.74

8.88

8.74

11.0410.4110.£4

DEG IS MAX HORIZONTAL SEPARATION IN WHOLE DE

- JEA.dL13iL~ibJ~S+/i2 -_ __ _g r>1wsZ~t TD -HSBBG.RE-AT____--~ .6PR RI i i ~blTAfeol~

Li�Z

- . I , ;- T --

-r-- i- I - t-i-t i i r .__ ._

i i L .

._. - . t-- M* - -; ' -l (

W Y tt-......._ ___-. _ 8 .7 --

BFE 10xl -

.- - -a --- L

I t ~ ~ 4 I

I i ! !

Pi1yTTTXh<.I ! -I -T 1_ _ 1 _, _ +L = --Ii-1

- i , .! := I II T i-U'

I - --v--,IZp-- i -4 '--,fix I--'-

_ -- -i, rI 1-' -+ t--

IE-:r 17.uVI /

�JVS5 �r- �

- 3A�A�

-:It1!)r-iF =

,--I i i i . 4 _ 1.,_1i

--- �

.. I_ 4-- -- 4 ,-

- as V l ,

IA4 M7

T- -- l-- 4, ~ -4 -- 'T ! I

I . 1

E T .I _ I0

== 1 I _

. i 1-7 , - ~~~~~~~~~~ I 1 ; I I . I ~~~~~~~~~~~ I I . ~ ._

:_ r_- , - -- -T 'T+,-__ --_7

I.t-- 1 : i .! J _ I y]i-I-i I1

:!i

".. L

,r-- a -- �- I-

7-

; --- . , --- . -. ,;. -,

. -.4 -7P? 2

1p 1; _

7 I / 1 T-_- -. I . I I q . - , t ! I I---- -- l: . __ _ ,, _ _.

_ - I _ - f j i ,

C/to-L-- e - LI.-Dt--*--~-- -

fi9T&Xcrck t 1--0--

=

= + -7 -T T _

Iu

. I I I I : , ,--F

-~ J !i!1 -----

. S r - f i ;,rIt l

_ .l

r : X : - - I- _ . 7 _., iF~_ V U Iffl~

Li

t

I

I

-

I

I

L.

- i : I hn

I

_ -I-r-- -

4 '-tt- 4-4 , - - - I - . .

_t j .

I __l ' I , :

t__ . , _ . .

i I_ _ _

___. - I !_0

__ _ ' _t __

- =

----3--t----- ---- lj

L_ i .t I I ;2

f ! | ii -

Lr iI iIF -- ----

-.- i � ; - -- i �

' 1---25. - -, ArI � � i

-- 7 r---r---_ -- Tj r - :i , :--T--- ~-= __ __ -+= i _ I >:--

1 ! i' I

~- ...... ."i

i ; 7 2 I_.+ . .. _ _ __ _; . q _

.

_ _ _ _ . E _

_ _ _

0 _ . . .:_ . ,, '. .,---- ; �--T-+-. _ . .. _ _ . ._ ; j _

.. _

_2 i , Ii , _ .1;1 ]i ,

6-7tt----J-L-L) ri !

-- -,- 4 , -. ---- i

Y-g ---j ---| --r . - -

- K - *- qiIc,!AFr -R/uoBFPI- LOIA 20 X 20 TO INCH

- .-- o -- Il CR fT - o H HAND LINE ROGRESIVELY ACCENTED...R400-~~~~~~~~~~~:~.~,~,~G~ESV' ,CCOCS-A

/MeAN SPEED DIFrFERENCE (/m/s$C)

t3EW~6X COLOCT1SEIj WINOC S-NSoIl (A8ve s°o& MIR)

Iv1AX. SEPIMfropJ hi I Honr(

M A1 M v" ;J A 0 D 7

1801979r IM M /' Q" a: 4

- - - - ~ eP r-jA

:~ - ~ FPI. OAM-20 X 20 TO 1 iNCH

-o -- o a- AI RC -it AF 7r 66 ., S o', ,,:10TH AND 20TH U.NE PROGRESSIVELY ACCEN [ED

~--0 y e- R4 Q&-Gots-o

MIF)N '13FPEEZ b1PF--CRtNCtS (tPI/Sc)

BEbtWCEIV C A. LQC vJrE N wiPJL) SENS5 (Af"l to S&01 /F)

AJAX. SCPnJIaT/N JP 2HOU(J

J .n .L-1980

.I- I" Ai M i M A fA z #9 S

._ - iA1RQAFT-RA68DF' TLCM-20 X 20 10'1 NCH

-- ~--- -, - CAlt cl Fl" -J--S./1 5TH, IOTHANO 20TH LINE FROGRESSIV'CLi ACCENTED

-X.'^ RHO I _ %,-T

MEAN RPEED PIPF-cRENCES (MI/SFC)

BrTWVJN COLOC¢Are-b WINb ShNSORS (iaOVE Soo *Ai)

MAX. CPtAPRT1o'J 3° 1 :-IoUIo i2,-.,

ADJ F

±980t m A /M ,4

4-

MK A A19'91 9 9

K s o

_ - - - _ . .. . . . . _ . . . . 11. IC, 20 X 20 "i) I NCH_ _ G- W- AIR cabs f } i, ah I S TCI, IOTi AND ,XI0 HINE iRC.GRE.5IOII I ACMENIED

SE --- x --x ---- Ret~-J'A

'WwCIN COLOCAT-'r-b W.IN> SM4.~SOjS (l/qVc Soo

MAX. SEPA/Zb&TlOr 30) I H-Ou

f FT~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

. [ }-!.jt. i;~iii':/, i 1: I ,:~ !iJ

i~~~~~~~~i d I ti iif 74ij+

' 'p

24744 L~~~~~~~~i I_ 4

±tiE tl Iifi~

*~~~~ I H, ~~~~~~~~ ~ ~ ~~ I~~~~ P . 41 *4~~~~~~~~~~~~~I4~.. 41' I~H144rj717 '~~~~~~~v < A 41I t-H i IH I I

jj~bi r~j~p42 ~ I I 4. .2 ~ tj4 . 11 I:. . .l

inf~~t~jj4~~. 4;; ~~~ L~i t~~$1. *~~~ V 1I i f I II - PiAfl 11t < . . 1 4

-41 _ T Jul. I K14 tl1d41,

)nff, ~ lan t

MI 1, ~ ~ 1

lil!T li

F Sa'1 8

"+?.i

Vin SPEE D VPIF'P'ENCE$ (vI/sec)AiRCl0AFT - ht a

Zr

AI ''o .; ; I Vrf) t '~;? 8 1 - '* 4 '8i

'~~~~~~~~~~~~~~~~~~~~~~~~~~~........."A...... ... ... . " ;

0

-* * 1- Alkte~2r-T-JPAoj3 - ,- - I.cRCT .-- R/-tozs3

.- I- x -- , K~a 11 O A - S A-'tMtE/AN SPE $ F ) DIFFERENLCE (M/.Sc)AMflX SEP/JICAfT9A/ot 3* / HVAu/

A4& Viz 60 P11

LI l

0 /N

lo

m A :m Z, A

WUWY_ ___* PM iv S V4; JAWU IA" -. - ___ at SW Tar AV __ yaw am - sJ4~~~

..........~~~~~~~~~~~~~~~~....... ...... ... j :~ .i ~..t? ,-- . ..... ,r !i

; 4- i . , , · L t i. ' '.

I. 'T T-Z~4--t1~ ----i-I- -Tt-t -

_ ... ,~~~~~~ jj4 .4

+ r T~~~~~~~~~~

;;;h i ~~~~~~~~~~~~~~~~~~~~f ' ....4~~~~~~~~2.~

, ~. ~ ... ,. ~.~.~:-'~ ~l ~.~~~~~~~~~~~~~~~~~~T- S o

I- ~ ~ ~ ~ ~ ~ ~ ~ ~~-~~~~~-rt~~~~~~ -11 -- - - -------

Tn4.t-f'':' ! ;!nfN ra " iq4i~~ii~Hir-I ~L-T i- T.] T{{ [44 2I71{

1:: :. II .: :

. . . . . . . I

. . .. . . . . . ...........,q , ,, ,4 : .. ,.....

. �7- ' --

.1i - ~ .. .... . I i &

v v

. . ,. A .:.:: :'. ---

-- A.'- ~-: .- 1:~::I !

LIha

....... ... ~ ~ ~ ~ .

- -,-, v,~~~~4 T

VI~~~~~~~~~~~~~~~~~~~~~~~

. . . . . .. . . . . .

1, :.. .X",. t o . LI Ift

l -r

- - -I iJI. . . . .

S .1 ....... I 7:7--e i --,A -� -i4 , -.- -, , , -, i.... .

.. . .. . I- tz

I

Ir

I

i

P

I

i

I

.

: S-

t.

I

t

Im L

1i

I

I �, :-.I . . . -

.I ., , , . ., , , , ,

.....- . ....... ...I - 1 . I........... - .......... ..........- - ......... -

T . . I. - . . L .

.- - - - - - . , . . .

. ..... I.... -. I..... ...... ....... ..

m:-�- -:: : : - .

,. . . f . � . +

� : :/ I : : -. ,LL Z+: : : - - .

:/ ...... I I

.... -

........................ 1. ........

- - - . - - - 4I I I .

- -. - . ....

I. -- , -

K. . I . ! 1 -. . . . 1 , :.I .

I'

EA

. . . . . . . .. . . . . f.. . . . . . . . I'a

_-:: , . I

0

AMhr/JA.(L SENSor Vs 8 ABOVE lCL4 5bt f- (,3', S/HcAJ4epEl;/7IoN)

~. _ _ .......A' -~. . . .....

--- G ; eS i:'., " ' ' " . '.0............... '...... ..;'. .,__- i : ' :. ....... -...... ' :.....i0' .90 ----a --r/qC~wr .. ..s - . . . _ -- 1 - . : - :

A - x fI : G:: -:

.- I

I... :: ::.,

* I .. .::.~~ ... I I;

1. - I,S.1

... . ..7 ..>~~~~~ V~~~ . a . .

50;

1 - : R. . ...,:: ::, ., _

I -I------ ", ---I - -_ =-I N ' . i

- -,,::

· . .~, ..I...,!.

I -.

./ :/..- ,;I..I., . t . . . .

1-: :: . , ,

I.I.....

I..I. 1 1... ........... '.. . . .... . . ... . . I:.

: \ . .

. .. .

I ' ''\'. .. I..

\:-... .,:I ........l ,\ I..

I V.. I.III 4 : . .

. h ; f o . . -

. .I + k. -

i- .. ,

t -~ ', 1 i.

l . ,4 I.. .7 . .., I ...... .

I

-O . . .P _. . ... F ( __ . . ._

NOV airc X4AM AV FEB

1. .. . . . . ..... I . .... . .... ....... I I ,

I' ' ' ''!-?It !.I .

I . t t+tT ,

V.

t-t............. - 1-7: 0 : :-: : :-;:: :1:- t. : ..-. i..;...... .: ! ....................."W~~ ~~~~~~~ :. r ---------- :: : -1I . . .. . . - .:: : :: : i~~~~~tt .

30 - :---

0 FIo F.B

LiEC N-9 - M O FCB

,, ....... .. I

t ,.. : ~ I - - 0- --

F

! -:. ! .; : :- - ! I . . . . . .

..- - - --.. . I

.~~~~~~~~~ .

S AR AMA ,

.. ; ......-! i '! !

| I .,

H �\

i i;, : : : !?. f I.

: 1 :.

i ... . .

.,i ;. ,.j . . .

r I .:I I..; I

F- -

.: :

.... 1

I... . .P.. m..

... ...... . .. . . .

::::::::: .. . .. ...

lzr J~~~x sUB Arl.. tr nr_

fAR

70 --

A~

I a

4 4 4 + 6 *I . . . . . . .g S - -~~~~~~~~~~~~~~~~~~~~~~

I . ' I ' ' -I ' I ' ' i. ' '' ' ' i * -- ''i I I I i .. t. .. I - _; -I:

.. I -- i : : j~ .. .. . . : . I I I -- I. .f. .~ I .. * . . -* |. .

_ : _ _4

,, ,_!.dt] T$ .. I ... __ --.. .II .. S: ',

- J '.-I

W%. F-

1

I

I

0

I

I

I

I

I

t

,

I1;

I

I

,

l

.: I ?. I . : - j. I.:. . ....,

I ~~~ . . .. .

-

I

!

I

iI

l

I

f

t_

l

I

d

I

i

l

/

I1

1

iI

l

I

....

. .:. : ., : : e

. . 0 .

. . . . . a a

. I ! . .

. . .6 z ~~~~~~~. . . . .

. ; , . ! . ... .At I.. .... ..A. :::

. . . . .

. I. . .. . I .

e . I.

:::. . ::.::.I I. .... i. ..... rI,.i.:...... ~ !..

9-

........ ~I . . . ..

.. . . ..I.: :.::1. .

;::::;?. .. . . . . : I

. i1 : -

.. . . .. ..; . . .: :

,:::: . .: -.-

7: .:.:::::

. . . . '. i I . . ; .: :

.I .; . ..

. ... . ...

:;, i:: '.

::~ :x f!;::: 1 :.. ,,,

...,-

.. NI :::/ ~ -11 :'::': A........

.1.1.............. s l .

r .... 4*w 7 _ ,1 t . . ..' . ' . .

: : : ., L. i�.:

! -! '-i' ' !t .

. . i1 . . . . ..

'. ' : . :

I .:.:.:

I . .. .

1. .

. - . .,i. . . I

::: i:..

. . . _, I, .

: , i- ! L

I .., .:j.

.,.. I '......

.. ..... ..:........ I . .:I . . . .

I

: . : : 'i * :. :.... iL . .4-- .. i, .i _ ....

,, . i ; j, ..

3 4 * t-t-; .'

_ .. ' ..... ... ;

[T'r i~' ~ :::I

:. I; . . .- ....''::1::t:I": :-I :

; . i i

: .,At-

7-- --

-.. i . .. r!.. . .. .. ,i

.: .:I . .: . .I

1�-<... ± 4 _....___ . . 1.............I..IV...........\. . .... -!- _______ .____I '. ::: - Id i ' X a T .CAF- } r, Yf! .- 1 - - |

t .: ' : xX

L---F. :.- :zr

Aswi4: . 1GoE9-sAI' .'GOEX-' d -.T-SAT: ..

... _.. ...I . . .

I.... _ I

- 1 Je1 ....... I1:: .... :-'1'/'-.' ' ................ £ .+'' - ''' ,,· /:.

1

1 - , : I

, .. . ....

, . ...... I -

.-

DrcI%7 Z 3 9o FEB

I

I~

I#S , ; \

. .\

:; :: ..

I"-: : :/:::

. . _ . .

. . . .:.: . ..:I... ... ..1:::-'::"::

.. ... . .

.:' ' : : .''t1;s . 4'1t'::::: ::':t~.. :.i:.i iiI .. 4..

I . +.t- .......,.. 4 ;;I.,:,,, j,: -

, f ,X kH

.I I~fj ,., . ...

o 1 ... .

:....-i. '1:!. . .. !

. . i C..

-~ I?'

7.4.7.7 )

I ::.:: : /.!I ,,;,,J.-jXI .;:2-i .:: 1

4:.;,:7,.

.I j......L.I [ ' :.

..: -: 1 -----------. :-1-1- -- --- i- .1 : I ;; * I r-... f. rt',tJ,.,

.; , ' . ; i , ! '

,: . i f , i .I i if :. I

1 : . :.

i ' . . . . . I.

. . I. ",, . . .I . .\ . .1 . .:

r . .7 !. .1..... ;lit.... i..,, I

: ; I :, ;

I . . 0 [ . ..I.... +....~

:: ;:;.:.:I ..; . . . ;

I ... .II. ...

-ntlL Aril .. . e. .. .. . .. . . . ...Dtr _Wm t4o FEB

. . ....... . . . .

I .~.. I ..I . . . .

a

hA~A- I/~, '"?--:Jr

' : ~ . ~~~~~~~~~~~~~~~ I I I... ] ", ."··I .. .. . ..: .'~ ~ ~ ~~~~~ ~ ~~ ~ ~~~~~~~ ~ ~ \4 .* ' /l | | /. / ' .-i -/ IL \ .I ............................. < 1 A 1'i __ 1 .' ..-<......................... I - . ,4 -e,, -

,\~0/A, :

4I

. .__. I . . I . . . I .

I ........

|;, .' e . i RI 6 I . .: . .

I

�1*�'

I.:/I,. .:..:

I. :: .+\ L ....'"'.1 . :' -'1: / " ::'1 / " N:I. Ix

I.

::: a

ON

*fX

! i:: !.-,.. -i.....: ..... . ..

: : , , 7 .... -

F,9 9

.... .... :

....

. .. . .: . . _ II 2. .i | 'I : ' ! I' I

:' 1; . '1 i: . . ., -

. . _

... ' _. - __I -I.. I ',f

, _ _ __ .4 . . . _

· .,. I

. . . _. - j i --I'----- ; --- I- i , i . I I

IL

1

'ji

I

I

;

i

I

II

i

-

..:r

i,

l

'I

w

l

t

I

I

I

':: .:. :?' AI. . . . I. .

I . . I , : ., . . . . .

::: I I ..

.; ., :_ . ..-,. , I _ . . ..

:I~ . ., :..I-

. . .. . ...

.1 . ,. .i, I . . . .

. .. . Id \ .

1:: . ': . ;-..

, . . . . . .

I . . .. .. ..

, .. .. . .. .

. . . . . . . . I: : ' ' : . . .:

.... I

I . I . I . . ... ,

,.., .. Ij +I..I

. . . :., .

. . -. . . .

I . . . . . . . I

1 . -i .....

i . . . /

I .. . .. .I . .. .

.t . . .

I ':: :': :.:: I

:.:.. :-y

.. .. j ....,..... s.

. . . .\ . .i.

.. \... I

.. ,,, - z .. !

.I : : : ; ; Ij I

.;. ., . , j, .I . . I I ._ . .

I .i . 'i J ;

"v;....:_ ..'.. .......I,

, . . - II,... . I, . .. . I

I : - :.. .I . . ......-.1 . . . . .. . ... -..t -+- -'

.'s I "'- '"\.: . .a. .'-

I

. -. I

.. ]... \I7-4

: .: . .I

1 I .. .

I , ':-::; i ¥I ., 1, . . I .

.... '... ;'\.' .- -; l

1 : I ' )0:

'

ffi

I f :---

/ : : . ..

: -4 . . .

.. .... '.

', ..

' I . _. ..

I .. 7.. I

. .-\ .... . .. . '/I

, , w:::..:.... I...

........ r.

.II......::::::::: .

. . .. . I .

/F.

i' 'i :

I . . . . ..

. .I :

., . . _.. .. ..

' '' ....';;;';::

...... i|::: i-:: : ! |

.. .: -: , . I 8. ... ... . .. ..... I I4'

I . . .' I . ' .4

nf -N-4ov- M,.

_vS J9AQ8 BCLOWSOO0 MB (;s L3fi s/AOUA szP1i/l

-

. .. .*.; ' ..t ....

fvw

. , .

I ';_1 ..I- . . -i

. ., ..| : , ; e i :. :.I . .,, I. . ./

~-~:.Ii , I-,

:; ; :;.;I

I_._I

I_. .!

i

R MS E1 5('2b 5A'TELL ITE -V$5 A/ICF-,l F-T- /ABOVE: '500 .M8 - .< I Nu/ SEPM)£hTj

... ..f . ... .. .......... .... .. .. .. .. .. .. ~* ........ ."-..-,,,.. ... .. .. ..,.. ....... ..... .r -". ..... ..

I r'u ' % UCT 'NOV DEC Qf///g"$

/C2v)

Iiaa

94)

I "ln ', vJ oVN . J Wot_- Sr r

V � I . .0- � , L---- L -.-- � L.

I1 ~

' I l l _ { ' I ± ---II t

0 .. 1.

I s-'' .

.0.. ....., . -

1- . ...... ...:: '. : : 1 .- :,,

7 i-. . ' i, :. ._. .

_ .... . ...

I

1 =

V___

I ... .... ' - _ 1 -p -. 0 , 1 ,~~I . . .. . . I .

--

N 5I I ":, - -* i.

. ... ,.... 1:...

. . .'; I I

:/ t. 2. .. ,.4. ; .

P .I .- . . ,. . I- . .1 .... I

S. {; . ,,, 4 I

! :. ..I ..;: ' t :4 .I. ...! I . . . ;

I . . .;' ' i

I1, \

ir .- ... I!t ..... 1 . I _I______*1 .. .

t

i sii ~Ti.' I..IqiI~I

DEC 1979 *,H./79 FEA AMM

W.w . .: -.: ' '

::lU _ ..

qT

JoI . .

/

4

- I . .....

.--IIv-

1 -. I - D1

. r I . .

IFig9. II

ASA '' |OE.-A '

coal] -r| GoS&7

7* _ r I -1q t - | - /- I - \ I . . . . .. I I.- _ . ._ ; t - A; \ ; i | r 1-- iI .A 4 e 7 4 - I IIIIIIII I -.il~~~~~~~~~~~~~~~~~~~~~~~i' I..........,._...,.I~~~~~~~

I

I

0

i

f

r

-

1

_i

l

I

t

1

l

!

l

'I

I .......I:. . . . .

::::::--- .. . . I . ..I

I I. . . . . ..:.. . .. . I. .. ..... i

. .. .. .I...1..... ,,i

I.... .....

v .: .......,; . . . f . . i .

1: . . . : : i .

i ... . .. ..

I. . .I ': ;I

. . :; .. . i .. .. .. .

. ... . ...

. !.. I..I..

,... I- . .. . . .

I '. >- . ..

II... .I.I -

1..... . . 1

1- ::::::.1

I

T . . :. . . .76 ,:.::::::: :.-::--

:: : : . q-

\ -:-A

. ' -- L._-i...

/.100 ,,

! j - _ l~---

... .. -I....... 77..

. . ......-,

_ _I_._.. . ._

....-::-

,...... ..

a . i . I§ I. I.. .

;If ; t :.

I .,, ~. . . .p . , 1

II. _. : : ;'' ' .: _

.. .... : i, -. ,

. . .

...I . .. .. r

. I .. 1. i

c111-1�

I .' , , ,-, ...

I . I . ' .' I....:; . ;,.I .i .T. .. ..

1 . , ; . 1 I I I .! I ' .

: :

. ... .-. .

... .. ,

. . .

. . . .. . .

t i _. . I . . .

'.i ''.

I'1'' 1 .,, I . .... .

. l ; ,

l i i .., . .. .. .,, . . . .

1: 'i i ':: I 1 I

. .t i7 . . .

1:':'I:. .1

A ; mw JUL Ab StP ' O

. . . . . ..,. ............ Ix :-::\:::;::

: . :\:: :..::::.\ :::'... \ ....

: ; .: : \ :-: ' '

.I.. .1:i:;1:

· ;. T-.. ~ .

. . . L , . . i 6,

, j"t,;, j, (1

I -.._., .\:. : -. .

'. . 1 , ' '. I . ' i:! . I

i*:4:, ,

i , . I . iIir ji I:,I .. ..

- i i I ! i i

r r I .

I ::: 1:.'I -

' _ gI X

0

76

440-

40.A. . .....: . . . .: : : : :

o~4X& -

A,S,,AR .' 'I Loes.At :-_

ICots-,a.:t-. - ,I." i,1'-.I,1t J:

' . . -' y . I

. 1: -. . . . .V

- - . . .

T. i I I

L-I' t, t. . -.;Lii ];;

1, ...-.

-: ' ... ..

1 ::: :'

L. . ,'q .- . .

' - S . - _ _ _

_

I

.... -...I....I. . .:.. .:' I: . .,

* - .s .' t' . .

[ !i?'i 'N.- ...<i .i......

I

I

! .; .I.. , .4li. I. . . - i. ..

I I--

, ; : - - .:

LNl ' . I 1,

?-; : .I"T :. I, [&,;I

· t .... 1.:.

.,,; . . .

,'t t~

. . . .,

:;-r ~t I:.-., -i i!t

I

)0: ASPAR:.. .... ...! ..... I...

I. : : : ' I

.s......... b......I.. .. .. ..!4. j.

K. i ' .:.:

.' ...

' i t ?.

1 =_

I .r .A,: :

., ': I,* I I

'S*-;.'.';

: :.::: :::s . :: :.;: ;.:-.:. :::.:: ::.: ; ::: : :;:: : ... .::.:::,: :- .~.' 'tt .,\ t'il ,, : : : : ::

.. ; ............,. ..,. .,;;.. ................................ . i. .: ' *t ** ,,, ........................ I' t ' i. . . . . . . . . . :::::'"

1~~~ ~ ~~~~~~~~~~~~~~~ ..I : :1

Jo I _ .. _. I

m .rl.,12:

DEcIM9' IN Je FES R APR pM IbAL ' ^

:';!itt.

.;..- 1�.-..", ' ! ;ii !... ' t '. I

t T-I t + ' Il

t -

je

SEP acr �..eov PEC

.I ... ..... 1 ....... . :.. ..

1, ... r..

_. _ _ _rr .. ... . ' ..-,I .... .Mt,

" . . ' :

I

'1I

-l

l -

------ -

-

-I-- - -

i- -t ---.: .: fl -, t + 1/ 1 ^ : r _ .!: .; : .: l . l y. .. : ; __ _~~~~~~~~~~~~~~~~~~~~~~~... l_:_L i _.. o .......l ................ .--.---- ~ / -m- -! - f-'- .... , -i I --:- i' ..... I............. -'!, -. 07 i ~_,

~ i I i1 . .

-;. .- ..../..-....I. .... .I, I -t .............. -... 'l ....j-: _ -, . ..- / "! : 1 --1~~~~~~~~~~~ L

,.__ _--- _ , --.-- ; . ~ i I _

I

I

I

I

1

.

t

I

J

iI

; ._4,. ~.r-

f. .: . ."II;

..I .. , r I

. . . . A. .? IIi* . . :

*. ... . .

.-. I. ._- - 7-I

. . . . .

-. -I......:: .I.. ::

. i-o :: ,__

.-I : . . s .

: : : .: .: :

: . . I - .

:. I.:: . I * :: . ::..::..:: : . ...

- T: - --

. . .... I

:- ... :: I

. t. . . I I .

. . . . ! I . . .

; : ::-; 1 : . . . . I

: . . . . .........:::: ............ :..

::. .:;:'

:: ::-

I

I ..

:~~~~~~ . .

! .. :. . :::. .:;.:

.: - . ! i

i/ 1

~:wi:'-': '.;:.::' :.f "' '

:I. . I Wi ! .: :....... . . ..

::.:"'1I

::::: .'I-.l

, 0 1

,___ ... ...;.-i.

1t , .t:,--t _;-- = ;.: . .:-1:'' ' . 1

r . . .

.,. -- j .''

' . .' ' ' t t' ..- .

. . .14 t -

' . . 1 Ir:-

.. .. f. . .. .. . .. .

I : I - ...-.i ; , .. . I

I .. .. .. ..

I. . . .:. . .. I

I.:.:.: . :. I'~~~~~~~~~~~~~~~~~~~~~~~~ . . '. .'

. . .

.: I 1

:: :~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~: . .: : : : : 4 : : . : . 1 . . . . .

: . . . . : . . :

SEP

: .t i i, tI .' it-I

. I ..

i I

. I't: ! -~.',.', ;I..

. ..

' .1. .a ; .....

.... I .I .... 'I·-: . ,..m

P(ca CT ^ t/v

W TN-T~~~~~~~~~~~~~~~P OM- ID X IO TO I INCHW

IOTH LINE lIEAVY

E-T PA ES ST ELLI Th MEAF D/FFE ENCE (M¶sEcHN, /. HC111,)

0 0 -- --- -0 V!

'0 I .. 61 -4.§ F . 4:.

X-4 a 7, 9 W& IIS If I Xi - - I

36 -~~~~~~~~~~~~d

- - - - - -0 - - - --

+21 -P t- -IQ 4 I-

Jo -~~~~N

- . I i-i~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~.L:ZA AJ Al MAY lvd v zrqlv AO r%~~~~~- - ..I-.- i

1I ~ ~ 0 -0 -

s'fEEL LHNb.t i' M'atC (AC'sbS) vS VE9rlcAz 'ti~t- 3i4M iN, C./ScL/Lrt (Lit-Mr.): Jia-rA~teE 's-ItELL~t-kbuI CdJLOCMtiJNs J@G Ml

0 - j.,/ .

. . o P ·.·t

* ~~ ~ ~ ~~~~~~ : . *

· ~ ~~~ e*

f: P ~ ~~~~ *c 0 * ..

, D ~~* -*..

0 *o. o?*~ 2 *o ** *

0~ ~ ~ 0 0 0 3

* 0 0.....0 &,/£

. ... .3 · .2 ·* 0.0 .1 .34

00. 00 0 '0 .:.0:

: 4. ~~~~009 0 ;O' * ' * 3. .Ž' 0*0 2

~~~~~~~~~~~~~~~~~* 0 0 0 0 0 0 y .,~~~~~~~~~ ..

' ' ' . ,* 0, *:. h L o * 6 E X t C0-0 *

~~~·..: . j . · ·.< . e · * ....... *.....* 0 0* 0 0

00 0 *

..' -. tf,:' ...... :. .. :...: .;.... *,. ,. .

p e .... * *0

: ' :0 S 0 0 0; * * *-* 00 d~. *

0 *i . - :

- *e

*00 * * -0,

t-: . st- FIG. 14 *-0.

· *

C0

*