Pasquill-Gifford (PG scheme). These are functions of usual · PDF filepresented nomograms (later revised by himself), which related Pasquill ... GZ and demonstrated that PGT curves

Atmospheric stability in Athens, Greece,

during winter and summer

B.M. Synodinou* & H.D. Kambezidis^

Institute of Nuclear Technology and Radiation Protection,

Email: vana@zeus. int-rpnet.ariadne-t.gr

"Atmospheric Research Team, Institute of Meteorology andPhysics of the Atmospheric Environment, National

Greece. Email: [email protected]

Abstract

The present study estimates the atmospheric stability for the Athens area basedon Pasquill's formulation. The results are limited to examining the stabilitycategories during the months of January (mid-winter term) and July (mid-summer term). These stability classes are examined from a 44-year databaseconsisting of hourly values of wind speed and direction, solar insolation andcloud cover. This data is regularly recorded at the National Observatory ofAthens located near the city centre. The results of the work consist of diagramswith mean monthly stability class values during January and July in the period1953-1996 and the mean daily variation of these parameters for the same monthsand period. Discussion of how these stability categories influence air pollution inthe Athens area is made.

1 Introduction

Stability conditions are among the most important parameters that affectthe dynamic processes occurring in the surface boundary layer. They areimportant factors to taken into account in atmospheric pollution studies.The stability criteria mostly used are the stability category estimations by

Transactions on Ecology and the Environment vol 25, © 1998 WIT Press, www.witpress.com, ISSN 1743-3541

272 Development and Application of Computer Techniques to Environmental Studies

Pasquill-Gifford (PG scheme). These are functions of usual

meteorological parameters such as wind speed and direction, solarinsolation and cloud cover. The pilot work of Pasquill[l] for flat terrain

introduced a set of six stability classes (A-F) as function of only windspeed, solar insolation and cloud cover. Gifford[2] later modified theseclasses in a more usable form. Turner[3] further modified the PG schemeby introducing one more stability class in which estimation of netradiation from solar altitude, cloud cover and ceiling height was made.Later, Turner[4] extended the above work to explicitly express stabilityclass in a more useful form in direct relation to incoming total solarradiation. This scheme, known as the Pasquill-Gifford-Turner (PGT)scheme, has been commented and later reviewed by Pasquill[5] himself,Gifford[6], Hanna et al.[7], Sedefian and Benett[8], Turner[9], Agrawaland Kulkarni[10] and others. The PGT scheme has been the milestone ofthe works involving stability. Nevertheless, there have also been othermethods to determine stability categories in a different formulation ormodify the PGT scheme, incorporating recent knowledge about thesurface boundary layer behaviour. For instance, Golder[ll] developed

procedures to relate the stability classes to surface-layer-stabilityparameters and to directly incorporate surface roughness. Smith[12]presented nomograms (later revised by himself), which related Pasquillstability classes to wind speed and sensible heat flux taking into accountestimated surface roughness and wind speed at the reference height of 10m. Sutherland et al.[13] formulated a semi-empirical expression whichrelates Pasquill stability category to wind speed and sensible heat flux,through a modified form of the Kazanski-Monin parameter that accountsexplicitly for surface roughness. Davidson[14] has given a modified

power law representation of the PG coefficients giving an alternativecurve to PG graphs. Sedefian & Benett[15], Mitchell[16], Irwing[17],Coppalle et al.[18] compared various stability schemes not necessarilyproposing that of Pasquill for their specific applications. Draxler[19]experimentally evaluated the accuracy of various diffusion and stabilityschemes over Washington, D.C.; he found that PGT curves givecomparable results to other schemes. Modifications to PGT scheme arealso reported for other types of terrain. Hasse & Weber[20] worked onthe conversion of Pasquill stability categories over sea. Erbrink &Scholten[21] expressed a new formulation for land and coastal regions interms of PG stability class indicators. Atmospheric stability has also beenstudied in complex terrain by various authors (e.g. Hurley[22], Koracin &Enger[23], Synodinou et al.[24], Petrakis et al.[25] examined theapplication of PGT curves to urban environment and found that they are


Development and Application of Computer Techniques to Environmental Studies 273

still valid). Between the most recent works there is the one by

Venkatram[26] who reviewed and commented PGT scheme and extended

its curves to include results of surface-layer-similarity studies, extending

the work of Golder[ll]. He used meteorological parameters such as L(Monin-Obukhov length), u* (surface-friction velocity) and z<> (roughnesslength) which have to be calculated by other semi-empirical methods(Van Ulden & Holstlag[27]). He presented analytical expressions for onlyGZ and demonstrated that PGT curves represent a number of possiblevertical curves associated with a surface release. The fact that the most of

these modifications are not complete and the usage of non-standardmeasured meteorological parameters that could introduce more

complexity and uncertainties in the determination of stability classes,

make PGT method valid in the usual calculations of atmospheric stabilityconditions. The PGT method also represents the standard basis for

comparison, because most applications of local character imply use ofthis scheme.

There are not a lot of statistical analyses of stability classpersistence in the literature. Among them the following works can becited. Satyanarayana & Agrawal[28] presented the long-term pattern of

Pasquill stability categories in the N.E. region of India. Agrawal &Satyanarayana[29] analysed the long-term trend of percentage occurrence

of stability classes in the north of India. Viswanadham & Santosh[30]also studied the relation between atmospheric stability and parameters ofthe surface-boundary layer over S. India. The atmospheric stability inGreece, regarding the Athens basin, has been studied mainly during thelast years by Lalas et al.[31], Synodinou et al.[26], Petrakis et al.[27]. Awork by Amanatidis & Zerefos[32] can be cited for the Thessaloniki areain N. Greece.

Aim of this paper is to statistically analyse stability conditions inthe Athens basin during January and July, the typical months of winterand summer in Greece. A data set of 44-year hourly values of windspeed, solar radiation and cloud cover is used. In the present workPasquill stability criteria are considered; these criteria are modified insuch a way that the climatic conditions of Greece are taken in account.

2 Topography - Data collection

Athens is located in a basin of area about 450 km^ surrounded on threesides by hills and mountains and at the southern part by the sea withabout 3.5 million inhabitants. A great number of industrial activitieswithin Greater Athens Area (GAA), in combination with a high daily



traffic are the main reasons for the serious air quality problemsencountered in Athens; such air pollution episodes occur during calmperiods in spring and summer.

The meteorological data used in this study is measured at theNational Observatory of Athens (NOA, latitude 37° 58 N, longitude 23°43 E, 107 m a.m.s.l.). NOA is located on the top of the small hill ofNymphs near the city centre 5 km from the coast. The data set comprises44 years (1953-1996) of mean hourly values of wind speed and totalhorizontal solar radiation. Cloud cover is measured at NOA three times

per day (0800, 1400 and 2000 h LSI). These values are consideredconstant three hours before and three hours after the time of observation,

except for that at 2000 h, that is considered valid until 0400 h.

3 Methodology used

The calculation of the stability classes according to Pasquill criteria,using NOA's wind velocity, solar radiation and cloud cover data in thementioned period, followed a modified formulation proposed by

Synodinou & Varsamis[33]. The modification concerns the finer

subdivision of wind speed and solar radiation daytime intervalsconsidered in the original Pasquill scheme. According to PGT methodtwo values of stability class for the same interval can often be computed.This can sometimes produce dubious results, because in dispersioncalculations one has to be able to define one stability class per interval forfurther calculations. Moreover, the fact that high amounts of solarradiation are generally received in Greece throughout the year, and bigchanges in the incoming total solar radiation can be registered in shorttimes, especially during morning hours, dictated the need to furthersubdivide the solar radiation intervals of the Pasquill scheme. Further,subdivision of wind speed intervals was necessary since large values ofsolar radiation with small values of wind speed correspond to much moreunstable conditions and large values of wind speed with small values ofsolar radiation correspond to more stable conditions. In this wayincreased wind speeds and reduced solar radiation values correspond tomore stable conditions. The stability classification scheme used in thisstudy is shown in Table 1.

4 Results and discussion

Figure 1 shows the daily variation of the stability categories in Januaryand July. Figure 2 shows the diagram of the mean monthly variation of



Pasquill atmospheric stability categories for January and July. Note that

the stability class numbers on the y-axes in both Figures are decimals

because of the averaging procedure. The dashed lines in Figure 2 are best

Table 1the wirandCLA-F anrefer 1

Intervals

WS1WS2WS3

WS4WS5

WS6WSl:wsWS2: 2 <WS3: 2.5WS4: 3 <WS5: 4 <WS6: ws

Stability dass in January and July

N)

G) 4k

01

0)

. Determination oid speed in ms , Rthe cloud cover ii

3 referred as numeo wind speed and

Rl

AAB

BC

C

R2

AB

B

BC

D<2: ws < 2.5< ws<3: ws<4' ws<6>6n

5.5 —

.O

4.5 —

.0 —

3.5 —

.O

2.5 —

.0 —

1 *^I I

2 4

"modified stability. the total horizont;i octals. In this worals in the range 1-solar radiation ink

R3

BBB

CD

D

R4

BC

C

CD

D

Rl:R>R2: 50 >R3: 37.5R4: 25 >R5: 1 2.5

\ January

^ — "

6 8 10 12Time (hoi

class PGil solar rark the sta6. WS1-):rvals use

R5

DD

D

DD

D

50R>37.5>R>25R>12.5>R

/

J

I I

14 16irs LST)

T schermdiation irbility cafctfS6 andd in this \CL>4/8-

FE

DD

D

J

uly

I I

18 20

;; ws isi Jcm'jgoriesR1-R5work.CL<3/8-

FF

ED

D

r—

i

22 24

Figure 1. Mean daily variation of stability class during January (solidline) and July (dotted line) in GAA (1953-1996).



fits. It is seen that in January the atmosphere is slightly stable i.e.stabilities just over 4 with an exception for 1994 when mean stability isjust less than 4. In July the atmosphere becomes more unstable (stabilityless than 4), as anticipated. From the best-fitted (dotted) lines in bothmonths an almost 20-year cycle can be detected. This 20-year stability

variation may be attributed to the 22-year solar magnetic cycle that caninfluence some weather parameters.

4.4

4.3

4.2

4.1

4.0

3.9

3.8

3.7

3.6

3.5

195O 196O 1970 1980Time (years)

1990 2000

Figure 2. Average annual variation of stability class during January(upper solid line) and July (lower solid line) in GAA (1953-1996).

The daily variation of the mean atmospheric stability conditions inthe Athens area ranges from 2.7 to 5.5 in January and 1.9 to 5.5 in July,as seen from Figure 1. Nighttime stability is about 5.5 (very stableconditions) in both months. Daytime stability "drops" to low values of 2.7

in January and 1.9 in July, i.e. very unstable atmospheric conditions. Thetime of minimum is different for the two months. Very unstableconditions are reached during winter at around midday and at 0900h LSTduring summer. This is anticipated; in winter lower temperatures andhigher amounts of cloudiness result in less turbulent surface layer, whichcomes to its highest activity around midday. During summer monthshigher solar insolation and temperatures combined with clear skiesproduce high turbulence in the lower layers of the atmospheric boundarylayer as early as a couple of hours after sunrise. This situation explainsthe frequent air pollution episodes occuring during summertime. Unstableair in the lower parts of the troposphere gives rise to temperatureinversions, which trap air pollution below.



Neutral stability is achieved in the evening (1700-2000 h LSI in

the winter and 1900-2000 h LSI in the summer). This occurs as the

atmosphere turns from unstable to stable conditions. Knowledge of thetime of this event is very important to air pollution studies andsimulations; most of air pollution modellers consider neutral conditionsor near neutral ones throughout the day.

Few similar results to this study are reported in the internationalliterature. Coppalle et al.[18] give a relation of air pollution dispersionparameter relation to Pasquill stability values in urban areas;Satynarayana & Agrawal[28] give the frequency of occurrence ofPasquill stability classes in N.E. India; Agrawal & Kulkarni[10] compare

atmospheric stability as determined by various methods; finally Agrawal& Satyanarayana[29] give the diurnal and seasonal frequency of stabilityclass occurrence over N. India.

5 Conclusions

Atmospheric stability is a measurement of tendency to suppress orenhance existing turbulence, Atmospheric stability is related to both windshear and vertical temperature structure, but it is generally the latter thatis used as an indicator of the condition. Atmospheric stability governs the

dispersion of the air pollutants. For these reasons, the present studyattempted the estimation of the average atmospheric conditions inAthens, a city with air pollution problems, for January and July during aperiod of 44 years (1953-1996).

The analysis of this work revealed the influence of the 22-yearsolar magnetic cycle on atmospheric stability. Furthermore, neutral ornear-neutral conditions seem to occur in January in general, but certainlyin the examined period (except for January 1994 when weakly unstableconditions prevailed). On the contrary, unstable conditions are prevailingin July and generally during summertime.

The daily variation of the atmospheric conditions consists of highvalues (stable conditions) in the nighttime and low values (unstableconditions) in the daytime. Neutral or near-neutral conditions areexpected in the evening when the atmosphere passes from unstable tostable conditions.

References

[1] Pasquill, F., The estimation of the dispersion of windbomematerial, Meteorological Magazine, 90, pp. 33-53, 1961.



[2] Gifford, F.A., Use of routine meteorological observations forestimating atmospheric dispersion, Nuclear Safety, 2, pp.47-51,1961.

[3] Turner, D.B., A diffusion model for an urban area, J. AppliedMeteorology, 3, pp. 83-91, 1964.

[4] Turner, D.B., Workbook of atmospheric dispersion estimates,

USEPA, ASRL, Research Triangle Park, North Carolina, 1970.[5] Pasquill, F., Atmospheric diffusion, 2"** Edn. Wiley, New York, pp

429, 1974.[6] Gifford, F.A., Turbulent diffusion-typing schemes: A Review.,

Nuclear Safety, 17, pp. 68-86, 1976.

[7] Hanna, S.R., Briggs, G.A., Deardorff, J., Egan, B.A., Gifford, F.A.& Pasquill, F., AMS Workshop on stability classification schemesand Sigma Curves-Summary of recommendations, BulletinAmerican Meteorological Society, 58, pp. 1305-1309, 1977.

[8] Sedefian, L. & Benett, E., A comparison of turbulenceclassification schemes, Atmospheric Environment, 14, pp. 741-750,

1980.[9] Turner, B., The long lifetime of the dispersion methods of Pasquill

in U.S. regulatory air modelling, J. Applied Meteorology, 36, pp.

1016-1020, 1997.[10] Agrawal, K.M. & Kulkarni, V.S., Comparative study on

prevalance of atmospheric stability classes as determined by selectmethods, Indian J. Environmental Protection, 12, pp. 431-438,1992.

[11] Colder, D., Relations among stability parameters in the surfacelayer, Boundary-Layer Meteorology, 3, pp. 47-58, 1972.

[12] Smith, F.B., A scheme for estimating the vertical dispersion ofa plume from a source near ground level. Proc. 3"* Meeting ofthe Expert Panel on Air Pollution Modelling, NATO/CCMS

Report No. 14, 1972[13] Sutherland, R.A., Hansen, F.B. & Bach, W.D., A quantitative

method for estimating Pasquill stability class from windspeed andsensible heat flux density. Boundary-Layer Meteorology, 37, pp.

357-369, 1986.[14] Davidson, G.A., A modified power law representation of the

Pasquill-Gifford dispersion coefficients, J. Air and WasteManagement Association, 40, pp. 1146-, 1990.

[15] Sedefian, L. & Benett, E., 1980. A comparison of turbulenceclassification schemes, Atmospheric Environment, 14, pp. 741-750,1980.



[16] Mitchell, A.E. Jr., A comparison of short-term dispersion estimates

resulting from various atmospheric stability classification methods,Atmospheric Environment, 16, pp. 765-773, 1982.

[17] Irwing, J.S., Estimating plume dispersion- a comparison of severalsigma schemes, J. Climate Applied Meteorology, 22, pp. 92-114,1983.

[18] Coppalle, A., Paranthoen, P., Rosset, L., Delmas, V., Francois, A.& Hamida, M., Comparison between dispersion parametersdeduced from wind fluctuations and Pasquill values: application toa short-range urban pollution case, Int. J. Environmental Pollution,5, pp. 671-678, 1995.

[19] Draxler, R.R., Accuracy of various diffusion and stability schemes

over Washington, D.C., Atmospheric Environment, 21, pp. 491-499, 1987.

[20] Hasse, L. & Weber, H., On the conversion of Pasquill categories

for use over sea, Boundary-Layer Meteorology, 31, pp. 177-185.

[21] Erbrink, H.J. & Scholten, D.A., Atmospheric turbulence abovecoastal waters: determination of stability classes and a simple

model for offshore flow including advection and dissipation, J.Applied Meteorology, 34, pp. 2278-2293, 1995.

[22] Hurley, P.J., An evaluation of several turbulence schemes forprediction of mean and turbulent fields in complex terrain,

Boundary-Layer Meteorology, 83, pp. 43-73, 1997.[23] Koracin, D. & Enger, L., A numerical study of Boundary-layer

dynamics in a mountain valley Part 2: Observed and simulatedcharacteristics of atmospheric stability and local flows, Boundary-Layer Meteorology, 69, pp. 249-283, 1994.

[24] Synodinou, B.M., Petrakis, M., Kasomenos, P. & Lykoudis, S.,

Atmospheric stability and wind circulation in Athens, NuovoCimento, 1995.

[25] Petrakis, M., Kassomenos, P., Lykoudis, S. & Synodinou, B.M.,On the atmospheric stability in the Athens basin, Third Int. Conf.Air Pollution 95, Porto Carras, Greece, 26-29 September 1995.

[26] Venkatram, A., An examination of the Pasquill-Gifford-Turnerdispersion scheme, Atmospheric Environment, 30, pp. 1283-1290,1996.

[27] Van Ulden, A.P. & Holstag, A.A.M., Estimation of boundary layerparameters for diffusion applications, J. Climate AppliedMeteorology, 24, pp. 1196-1207, 1985.



[28] Satyanarayana, Y.V.V. & Agrawal, K.M., Long-term trend

analysis of Pasquill stability classes over north-east Indian, J.Environmental Protection, 11, pp. 502-506, 1991.

[29] Agrawal, K.M. & Satyanarayana, Y.V.V., Diurnal and seasonalfrequency of atmospheric stability classes in the planetaryboundary layer over northern India, Indian J. EnvironmentalProtection, 11, pp. 20-24, 1991.

[30] Viswanadham, D.V. & Santosh, K.R., On the relation betweenatmospheric stability, surface turbulence and mixing height overSouthern India. Theoretical Applied Climatology, 49, pp. 19-,1994.

[31] Lalas, D.P., Catsoulis, V. & Petrakis, M., On the consistency of

stability classification schemes when applied to non-homogeneousterrain, Atmospheric Environment 13, pp. 687-691, 1979.

[32] Amanatidis, G. & Zerefos, Ch., Pasquill stability classes

calculation for the region of Thessaloniki, Private Communication.[33] Synodinou, B.M & Varsamis G., Atmospheric stability category

determination in the Athens area, DEMO 88/4J, 1988.

[34] Synodinou, B.M. & Kambezidis, H.D., Atmospheric stabilitycategory in Athens during spring and autumn, 4 Hellenic Conf onMeteorology-Climatology-Atmospheric Physics, Athens,September, 1998 (accepted for presentation, in Greek).


Documents

Pasquill-Gifford (PG scheme). These are functions of usual · PDF filepresented nomograms (later revised by himself), which related Pasquill ... GZ and demonstrated that PGT curves