Fulltext Sunshine Duration
Content
Theor. Appl. Climatol. 37,150-157 (1986)
Theoretical and Ap.plied Climatoiogy
© by Springer-Verlag 1986 551.521.11(495)
Department of Meteorology and Climatology, Aristotelian University of Thessaloniki, Greece
A Study of the Daily and Hourly Change of Sunshine Duration Values on Crete island
A. A. Bloutsos and P. J. Pennas With 3 Figures Received February 23, 1985 Revised March 12, 1986
Summary
In this paper the daily and hourly values of sunshine duration, as well as the spells of virtually sunless days (daily sunshine duration ~< 3.0 hrs) in Crete island are examined for the decade 1971-1980, which has been chosen for homogeneity of the existing observations. We have found that the sunshine duration at meteorological stations on the south side of Crete island exceeds the sunshine duration at the meteorological stations located on the north side of the island, which is attributed to the combination of Crete's relief and the prevalence of the north winds. We have also found that the sunshine duration measured at the meteorological stations on the east side of the island during the cold period of the year (October-March) exceeds the sunshine measured at the meteorological stations on the west side. This excess is attributed to the influence of the moving cyclones of the Mediterranean Sea especially during this season. We also believe that the moving cyclones are responsible for the same picture which appears when we study the virtually sunless spells with 1,2 or three consecutive days at all the examined meteorological stations, while virtually sunless spells with more than 3 days show a concentration on meteorological stations on the north side of the island which is caused by the ground relief and the north-blowing winds. Finally, we found that virtually sunless spells follow a modified G. Polya distribution which, using the X2 -test, gives a very good fit at the 95% confidence level.
Zusammenfassung Eine Studie der t~lichen und stiindlichen ~aderung der Sonnenscheindauer auf Kreta
Es werden die t~igliche und sttindliche Sonnenscheindauer sowie Tage mit kurzer Besonnung (t~igliche Sonnenscheiffdauer ~< 3,0 Stunden) auf Kreta fOr die Dekade 1971-1980, die aufgrund der Homogenitat tier Beobachtungen ausgew~ihlt wurde, untersucht. Die Sonnenscheindauer auf der Stidseite der Insel iibersteigt die auf der Nordseite, was auf das Relief Kretas und die vorherrschenden Nordwinde zurtickgefiilart wird. In der kalten Jahreszeit (Oktober bis M~irz) tibersteigt die Sonnenscheindauer im Osten die im Westen. Dieser Unterschied dtirfte auf den zu dieser Zeit starken Einfluf~ der wandernden Zyklonen im Mittelmeer zurtickzufiihren sein. Es wird vermutet, dat~ das vermehrte Auftreten von ein bis drei aufeinanderfolgenden, praktisch sonnenfreien Tagen im Westen auf die gleiche Ursache zuriickzufiihren ist, w~ihrend mehr als drei sonnenfreie Tage konzentriert im Norden auftreten, wohl wegen des Reliefs und des Nordwinds. SchlieNich wird gezeigt, dat~ die praktisch sonnenfreien Tage elner modifzierten Polya-Verteilung folgen. Ein X2-Test ergibt eine gute Obereinstimmung mit einem 95% Konfidenzintervall. 1. Introduction Diffuse and direct incident solar radiation plays an i m p o r t a n t role in controlling the air tern-
Daily and Hourly Change of Sunshine Duration Values on Crete Island
2 4" oo' 25" oo'
151
26"00' 36"00" 35"30"
35o00 '
_
ilMBAK~ON~___, A P ~
q Fig. 1. Map of Crete showingthe relief
E TRA
3~'06
perature of the environment in which human activity is present and helps in the growth of animals and plants, which are also very important factors in ecological equilibrium. Greece is characterized on the one hand by a large dispersion of islands in the Aegean Sea in an area twice large as more than continental Greece and a lack of fossil fuels on the other, so the use of solar energy will help us to solve vital energy problems. Thus, the necessity of knowledge, on a climatological basis, o f the available solar energy is very clear. Unfortunately, the majority of existing meteorological stations do n o t measure solar radiation. Thus we try to estimate the incoming solar energy using indirect methods and such a method, possibly the most convenient, is to use the sunshine duration in our calculations. For this purpose many other scientists have also worked on this topic, among whom we must refer to Macris (1976), Karapiperis etal. (1974), Flocas (1980), Helmes and Jaenicke (1984). From the whole Greek area we have chosen to study the sunshine duration on Crete island because, in our opinion , it fulfills the following requirements: (a) It is the most southern and the largest island of the Creek territory and also one of the largest islands in the Mediterranean area. (b) The people on this island have mainly agricultural jobs and specialize in the production of early season vegetables. (c) Crete is o n e of the most well-developed Greek tourist areas. (d) Last but not least, Crete is covered by a satisfactory network of meteorological stations
measuring the sunshine duration using the same instrument (Campbell-Stokes sunshine recorder). The locations of these meteorological stations are shown on the map of Crete in Fig. 1.
2. Data and Method
The data on which this paper is based have been gathered from the archives of the National Meteorological Service (EMY) for the six meteorological stations of the island which contain records for the hourly sunshine duration for every day. The sunshine duration at all the meteorological stations is measured using Campbell-Stokes sunshine recorders (WMO 1971) which measure sunshine duration with an error of -+5 min. Obstructions which could cause errors in the measurement of sunshine duration are negligible at all the examined meteorological stations. The time period between 1971-1980 will be examined because of the completeness of data during this time. The daily sunshine duration is grouped into 6 classes, 0.0, 0.1-3.0, 3.1-6.0, 6.1-9.0, 9.1 - 12.0 and > 12.1 hrs per day. In this work we study the annual distribution of the mean hourly values for each time interval and also the mean monthly values for every meteoorological station separately. Moreover, when we consider as "sunless" those days for which the daily sunshine duration does not exceed 3.0 hrs, we calculate the number of 1, 2, 3 , . . . consecutive sunless days for the cold period of the year, i.e. October, November, December of one year and January, February and March
152
A.A. Bloutsos and P. J. Pennas
Table 1. Geographical Coordinates of the Meteorologieal Stations o f Crete Island and the Mean Daily Sunshine Duration for Every Month and Corresponding Standard Deviation for the Period 1971-1980 Khania = 35°30 ' X = 24°02 ' M Jan Feb Mar Apt May Jun Jul Aug Sep Oct Nov Dec (310) (283) (310) (300) (310) (300) (310) (310) (300) (310) (300) (310) 3.41 3.98 5.93 7.49 9.82 11.98 12.19 11.41 9.28 6.02 4.94 3.58 o 3.07 3.32 3.53 3.92 3.26 1.80 1.15 1.22 2.43 3.53 3.14 3.04 4.32 Suda ¢= 35°33 ' X = 24007 ' M 3.61 4.08 6.03 7.63 9.79 11.78 12.12 11.44 9.29 6.35 5.15 3.74 7.65 o 3.12 3.36 3.63 3.94 3.30 1.96 1.12 1.31 2.54 3.50 3.12 3.00 4.77 Tymbakian ~0= 35°04 ' X = 24°45 ' M 4.69 5.24 6.88 7.81 9.74 11.60 12.16 11.63 9.85 7.31 5.92 4.60 8.13 o 2.79 3.26 3.15 3.65 3.10 1.86 0.94 1.54 2.04 2.94 2.77 2.83 3.79 Iraklion ~o= 35020 ' X = 25°11' M 3.68 4.17 6.02 7.47 9.81 11.86 12.10 11.14 9.37 6.22 5.10 3.79 7.58 o 2.94 3.32 3.46 3.90 3.13 1.97 1.15 1.28 2.22 3.40 2.88 2.96 4.17 Ierapetra ~o= 35°00 ' X = 25°45 ' M 5.10 5.65 6.71 7.62 9.95 12.09 12.36 11.90 11.06 7.65 6.68 5.18 8.34 o 2.67 2.99 3.00 3.64 3.06 1.73 0.99 0.78 1.74 2.79 2155 2.84 3.65 Sitia ~0= 35°12 ' X = 26006 ' M 3.81 4.47 6.23 7.26 9.77
11.86 12.03
o 2~90 3.23 3.27 3.92 3.29
1.83 1.04
10.90 9.22 6.42 5.37 3.89 7.62
1.05 2.05 3.13 2.78 2.94 4.02
Annual(365.3)7.52
o f the following year for each meteorological station. For the winter 1970-1971 only the months January, February and March were used while for the winter 1980-1981 only the months October,, November and December were used. We must also note here that for some spells we calculate the number o f consecutive days using either the last days of September or the first days o f April. Finally we give a theoretical distribution for the nearly sunless spells for each meteorological station using a slightly modified Polya method (Arley et al. 1973) because we found that Polya's m e t h o d does not fit well to empirical data while the modified distribution, using the X2-test, fits better at the 95% confidence level. According to Polya's method the probability of having N ( x ) spells of x consecutive days is given by the equation
Px = P x - 1 ( m + ( x - - 1 ) d ) / x (1 + d ) ,
next one, E ni is the total number o f the consecutive days, S i = x - - 1, x = 1 , 2 , 3, . . The number o f the expected occurrence of an event among N trials is given b y the equation
v = P'N.
The modified Polya m e t h o d gives the number o f spells o f consecutive days using the formulas
2 m e -m
N(1) = k N " (l+d) N(2) = X
m+(x--1)d x(l +d)
•(rn + d)
m+d
(1)
a N(1 ) (2)
4(1 + d)
N(x
N(x) =
1),x=2,3,4,...(3)
where k, X are empirical parameter. The fit o f theoretical to empirical data is tested by using the ordinary ×2-test.
where Po = 1 / ( l + d ) m/a, m = ( N - - ~ n i ) / ~ n i is the expected number o f days (mean), 02 = ( ~ n i s 2 / ~ , n i ) - - m 2 is the variance, N is the total number o f days, d = (o 2/m) 1 is a parameter o f the influence Of an event on the
3. Discussion
3.1 General
Table 1 gives the name o f the meteorological station, its geographical coordinates, the mean
Daily and Hourly Change of Sunshine Duration Values on Crete Island daily sunshine duration values for every month with the corresponding standard deviation as well as the number of observations per month, common to every meteorological station, for the period 1971-1980. As we can see from the above table or as has been mentioned by other scientists, e.g. Pennas (1976), Crete island could be divided, regarding the geographical distribution of sunshine duration values, into two individual parts using an axis from West to East along the mountain crest (Fig. 1), i.e. into the north part and the s o u t h one. Thus, on an annual basis the excess in sunshine duration values of the south side o f the island is about 300 hrs/year while on a monthly basis this could rise up to 50 hrs/month. This difference between the two sides of the island is attributed to the more fregment formation of clouds on the slopes of the mountains exposed to the north winds, due to the forced upward motion of the prevalent north winds during the year in this area (WMO ! 964). But differences in sunshine duration values are also apparent between the meteorological stations of the same zone during the cold months, in that the values for the meteorological stations located on the east edge of eacl~ zone exceed those located on the west side of the same zone. The above differences during the cold period are to be attributed, in our opinion, more to the peculiarities of the atmospheric circulation above the studied area, Machairas (1983), Karoulias (1975) and less to the ground relief. In fact, during the cold period of the year, cyclones of the Mediterranean Sea traverse this area with a general direction from SW to NE and as a result they influence first and more strongly the west side o f thee island. All t h e differences mentioned above are statistically tested at the 95% confidence level and found to be significant.
153
From Fig, 2 we can see that t h e sunshine duration exceeds 30 hrs/month for every hour interval from 08:00 to 16:00 L T (06:00 to 14:00 GMT) for July and A u g u s t . Low sunshine duration values which are apparent during the winter months, especially in the morning and late afternoon hourly time intervals are mainly attributed to the short duration of the day. For the winter months, the sunshine :duration for each hour interval from 08:00 to 16:00 LT has such values that often exceed 10 hrs/month. The isoline of 25 hrs/month for each hour of the day, which correspond to a value of more than 80% of the maximum possible sunshine duration, is between 08:00 to 16:00 LT during the months from May to September for all the examined meteorological stations, while for the meteorological stations located on the east side of the island, these isolines are extended till October.
3.3 Daily Values
Grouping the 3653 daily sunshine duration values for every meteorological station for the decade 1971-1980 into six classes according to the limits mentioned at the beginning of this article, and calculating percentages of occurrence, we draw Fig. 3 (a to f). When we study this figure we can see that, if we exclude the completely sunless days and the days where the sunshine duration is greater than 12 hrs/day, there is not any sufficient difference between the east and west side of the island. But when we study only the completely sunless days (sunshine duration 0.0 hrs/day) we can see that the percentage of their appearance on the meteorological stations of the west side is sometimes twice to three times greater than the percentage for the meteorological stations on the east side of the island.
3.2 Study o f Hourly Values
The mean values of sunshine duration in units of hours per month for each hour of the day and for every month have been calculated from our analytical records and are shown in Fig. 2 (a to f) where for every meteorological station isolines of sunshine duration appear.
3.4 Sunless Spells
Here we are going to examine the nearly sunless spells only during the cold period of the year. As a nearly sunless day we consider a day where the sunshine duration does not exceed 3.0 hrs/ day. The total number of such days for any meteorological station are about 25% of the
154
A.A. Bloutsos and P. J. Pennas: Daily and Hourly Change of Sunshine Duration Values on crete Island
a hrsI . . . . . . . . . . . .
b hrs 1
. . . . . . . . . . .
,, -°\\\\,
o..-
,.! -.:o.X.\-x\
k._)
;/'/o~
hrs/
;
:
:
:
,
,
,
,
,
,
,
i
I0,
o
o
I
e d
~ ~, i
,~ J
J
~
~
8
.
~
J / ~ i
~ ,I -I k ~"6
N~
hr: . . . . . . . . . . . .
h!s f
. . . . . .
Fig. 2. (a-f) Mean monthly sunshine duration (hrs) for each meteorological station
a
1 F F M
b
i
IJ:,~,>~,:~:,:~,~:,;.~l
,;;~, :.'.~,,~7~b~:!'~ !~:7'/~.~"::i ~' ;3 ~' ~.7~:-,,7 ~' "
I: i/Af/llllllltlll.*l~ .
KHANIA
]
i
om ~ ~ ~ ' ; , ) ~ 7 ~ j
.
'!~i:!'~':~.~x,':S'.'-,",~'~Ti, ~:[~./~"); "~'~:,~ ~:~;"."V~
(':.,t
.~:,'-.~,,~'. ,;177
~..;~ .,t:.' D:A~:. ,.t .- 4 :..~,~. ~-~,:~,:,:; ~D',';u
o
mm~)iii!,)){!i ii))))ii~: ~ ; ! 1 :7:~{i ~ ~!;:
oC
io
~
loo;/o
d
ii
~i !!
~ ,~
" i!.
,t~:
x:( '~,-'. ~:.>.:: 7 "c~ )
o
0
~ii:
20 60
/
100% 0 20 60 100%
e
f
l)))iiiii!)i!)iiii)))i))))))i~!l
~! i "~ ~ ~ ~L~;~~,~ ' ~t~+"h.
I ~,~'. 7 "i' "> .")"c2f, r 7Y. L~'~ 1 ~. ;:'D :, D:;:~,., ;:,~, x,, A,~,..I
W. ,~ ~.~
"
I
TM A I N I B KO ]
~ ~{!ii~'7)i;~;l
B, ;, ;,~.s ,v, P,'~-'~'~ ~v :, v :,Z'l
IRPT7 E A ER A
IJ "..r/il/I/I/Ij[fT,.,,G, ~3~+i,~y<')/~'.:3',.::.'!-,~::L~,y ..~ ~ 7' 7.~% ;tt~L,7..i'7 / ~ ..;7:.' :~i
° 17
I ~ / ~
;iil!:~71!}:i!'iTi;~i:i!;!i~il
°
o
m
0,0 0.I-30
NNN
m
).I-60 6.1-9.0 g.l-12J) ~-1211
Fig. 3. (a-f) Percentage of occasions of daily sunshine totals for each meteorological station
156
A.A. Bloutsos and P. J. Pennas
Table 2. Observed and Expected Spells of Consecutive Actually Sunless Days (sunshine duration ~< 3.0 hr/day)as Well as Coefficients and Chi-Square Value for Each Meteorological Station 1 Khania Suda Iraklion Sitia obs exp obs exp obs exp obs exp 2 3 4 5 6 9 8 6 8 7 4 6 4 6 4 5 0 3 1 1 8 9 10 11 12 13 14 15 16 17 2 3 0 4 3 2 2 5 1 4 3 2 2 0 .1 3 2 1 5 2 1 2 1 1 1 1 0 0 2 1 2 1 0 1 0 1 0 1 0 1 1 1 1 1 0 0 0 i 0 0 0 0 0 0 1 0 0 0 1 0 0 0 0 0 1 0 0 .0 1 0 K X Chi-square value 16.263 12.011
150 53 29 14 14 130 47 29 19 12 150 53 29 12 12 140 50 30 19 13
1.5 0.333 1.5 0.333
159 60 39 8 13 10 182 63 36 21 13 8 166 50 30 15 15 135 46 25 14 8 179 65 20 4 198 51 22 10 166 51 12 134 70 17 4 4 4 4 1 1 5 5 0 2
1.5 0 . 3 3 3 32.076 1 0.33 1 0.333 25.024 14.745 14.308
Tymbakion obs exp Ierapetra obs exp
0.5 1
total o f e x a m i n e d days o f the cold period of the decade 1971-1980. The number o f nearly sunless days f o r every meteorological station with 1, 2, 3 etc. consecutive days are given in Table 2. As we can see from this table, the meteorological stations on the south side of Crete island (Tymbakion and Ierapetra) do not have more than one spell o f 9 consecutive nearly sunless days while the meteorological stations on the n o r t h side have at least one spell with 10 or more consecutive days, up to one spell with l7 consecutive days. Another difference between the east and west side o f the island is also apparent in the m a x i m u m number of consecutive sunless days where longer and more spells are observed on the west side. If we omit the spells with more than 3 consecutive sunless days then all the meteorological stations have an almost equal number of spells and there is no difference either between any o f the meteorological stations. We believe t h a t the uniform distribution of the spells with one or two consecutive days is to be attributed to a factor common to all the meteorological stations and this factor we consider to be the influence of moving cyclones which over the examined area have a direction from SW to NE.
On the other hand, the difference in the number o f spells with more than 3 nearly sunless days can be attributed, we believe, to a combination of the ground relief and the prevalence o f n o r t h winds. Using the eqs. (1), (2) and (3) with appropriate k and X coefficient values, given also in Table 2, o f the modified Polya method, the theoretical expected numbers o f spells for meteorological stations are given again in Table 2, while the goodness o f fit is tested using ×2-test at the 95% CL. The results of the above test show that the fit should be considered satisfactory for all the meteorological stations except Iraklion.
4. Conclusions Concluding at this point what we have previously analytically discussed, we could say that to study the sunshine duration, Crete island should be divided into two parts: the n o r t h side which is characterized by a shorter sunshine duration on a daily, m o n t h l y or annual basis and also with longer nearly sunless spells as compared to the south side which is characterized by longer sunshine duration, fewer and shorter spells with sunless days.
Daily and Hourly Change of Sunshine Duration Values on Crete Island There could be a second division o f the island in to the east and west if we consider the m o n t h l y sunshine duration during the cold -period ( O c t o b e r - M a r c h ) where the east part o f the island exceeds that o f the west. This difference between the two parts cannot be observed either during the warm period or in an annual comparison. The primary reason for the difference o b served is the ground relief combined w i t h the prevailing upslope. North winds, over the northern part cause the formation o f clouds while t h e same winds, having downslope characteristics over the southern part help to dissolve clouds.
157
Acknowledgement
The authors wish to thank Dr. Ch. Sahsamanoglou for his very valuable suggestions on this work.
Flocas AA (1980) Estimation and prediction of global solar radiation over Greece. Solar Energy 2 4 : 6 3 - 7 0 Helmes L, Jaenicke R (1984) Experimental verification of the determination of atmospheric turbidity for sunshine recorders. J Clim Appl Met 23:1350-1353 Karapiperis L, Katsoulis B, Papachristopoulos C (1974) Contribution to the study of sunshine duration in Greece. Mere NOA Set II, No 3 8 Karoulias A (1975) The Saharian depressions. Ph D thesis, Univ of Thessaloniki, Greece, in Greek Macris G (1976) On the distribution of solar energy in Greece. Memoirs of the National Observatory of Athens, Ser II. Meteorology 43 Machairas P (1983) Les types des temps d~pressionnaires perturb~s au-dessus de lamer Eg6e. Riv Met Aeronautica XLIII: 13-22 Meteorological Office (1964) Weather in the Mediterranean, vol I, 2nd ed, HMSO, pp 7,119 Pennas P (1976) Sunshine duration in Crete. Sci Ann Fac Phys and Math, Univ of Thessaloniki, 16, 357 WMO (1971) Guide to meteorological instruments and observing practices, no 8 TP3, pp 32-35
References
Arley R, Grisollent H, Guilmet B (1973)Climatologie. M6thodes et pratiques, pp 136-138, 179
Authors' address: Dr. A, Bloutsos and Dr. P. Pennas, Aristotelian University of Thessaloniki, Department of Geology, Sect. Meteorology-Climatology, Thessaloniki 54006, Greece.
Theoretical and Ap.plied Climatoiogy
© by Springer-Verlag 1986 551.521.11(495)
Department of Meteorology and Climatology, Aristotelian University of Thessaloniki, Greece
A Study of the Daily and Hourly Change of Sunshine Duration Values on Crete island
A. A. Bloutsos and P. J. Pennas With 3 Figures Received February 23, 1985 Revised March 12, 1986
Summary
In this paper the daily and hourly values of sunshine duration, as well as the spells of virtually sunless days (daily sunshine duration ~< 3.0 hrs) in Crete island are examined for the decade 1971-1980, which has been chosen for homogeneity of the existing observations. We have found that the sunshine duration at meteorological stations on the south side of Crete island exceeds the sunshine duration at the meteorological stations located on the north side of the island, which is attributed to the combination of Crete's relief and the prevalence of the north winds. We have also found that the sunshine duration measured at the meteorological stations on the east side of the island during the cold period of the year (October-March) exceeds the sunshine measured at the meteorological stations on the west side. This excess is attributed to the influence of the moving cyclones of the Mediterranean Sea especially during this season. We also believe that the moving cyclones are responsible for the same picture which appears when we study the virtually sunless spells with 1,2 or three consecutive days at all the examined meteorological stations, while virtually sunless spells with more than 3 days show a concentration on meteorological stations on the north side of the island which is caused by the ground relief and the north-blowing winds. Finally, we found that virtually sunless spells follow a modified G. Polya distribution which, using the X2 -test, gives a very good fit at the 95% confidence level.
Zusammenfassung Eine Studie der t~lichen und stiindlichen ~aderung der Sonnenscheindauer auf Kreta
Es werden die t~igliche und sttindliche Sonnenscheindauer sowie Tage mit kurzer Besonnung (t~igliche Sonnenscheiffdauer ~< 3,0 Stunden) auf Kreta fOr die Dekade 1971-1980, die aufgrund der Homogenitat tier Beobachtungen ausgew~ihlt wurde, untersucht. Die Sonnenscheindauer auf der Stidseite der Insel iibersteigt die auf der Nordseite, was auf das Relief Kretas und die vorherrschenden Nordwinde zurtickgefiilart wird. In der kalten Jahreszeit (Oktober bis M~irz) tibersteigt die Sonnenscheindauer im Osten die im Westen. Dieser Unterschied dtirfte auf den zu dieser Zeit starken Einfluf~ der wandernden Zyklonen im Mittelmeer zurtickzufiihren sein. Es wird vermutet, dat~ das vermehrte Auftreten von ein bis drei aufeinanderfolgenden, praktisch sonnenfreien Tagen im Westen auf die gleiche Ursache zuriickzufiihren ist, w~ihrend mehr als drei sonnenfreie Tage konzentriert im Norden auftreten, wohl wegen des Reliefs und des Nordwinds. SchlieNich wird gezeigt, dat~ die praktisch sonnenfreien Tage elner modifzierten Polya-Verteilung folgen. Ein X2-Test ergibt eine gute Obereinstimmung mit einem 95% Konfidenzintervall. 1. Introduction Diffuse and direct incident solar radiation plays an i m p o r t a n t role in controlling the air tern-
Daily and Hourly Change of Sunshine Duration Values on Crete Island
2 4" oo' 25" oo'
151
26"00' 36"00" 35"30"
35o00 '
_
ilMBAK~ON~___, A P ~
q Fig. 1. Map of Crete showingthe relief
E TRA
3~'06
perature of the environment in which human activity is present and helps in the growth of animals and plants, which are also very important factors in ecological equilibrium. Greece is characterized on the one hand by a large dispersion of islands in the Aegean Sea in an area twice large as more than continental Greece and a lack of fossil fuels on the other, so the use of solar energy will help us to solve vital energy problems. Thus, the necessity of knowledge, on a climatological basis, o f the available solar energy is very clear. Unfortunately, the majority of existing meteorological stations do n o t measure solar radiation. Thus we try to estimate the incoming solar energy using indirect methods and such a method, possibly the most convenient, is to use the sunshine duration in our calculations. For this purpose many other scientists have also worked on this topic, among whom we must refer to Macris (1976), Karapiperis etal. (1974), Flocas (1980), Helmes and Jaenicke (1984). From the whole Greek area we have chosen to study the sunshine duration on Crete island because, in our opinion , it fulfills the following requirements: (a) It is the most southern and the largest island of the Creek territory and also one of the largest islands in the Mediterranean area. (b) The people on this island have mainly agricultural jobs and specialize in the production of early season vegetables. (c) Crete is o n e of the most well-developed Greek tourist areas. (d) Last but not least, Crete is covered by a satisfactory network of meteorological stations
measuring the sunshine duration using the same instrument (Campbell-Stokes sunshine recorder). The locations of these meteorological stations are shown on the map of Crete in Fig. 1.
2. Data and Method
The data on which this paper is based have been gathered from the archives of the National Meteorological Service (EMY) for the six meteorological stations of the island which contain records for the hourly sunshine duration for every day. The sunshine duration at all the meteorological stations is measured using Campbell-Stokes sunshine recorders (WMO 1971) which measure sunshine duration with an error of -+5 min. Obstructions which could cause errors in the measurement of sunshine duration are negligible at all the examined meteorological stations. The time period between 1971-1980 will be examined because of the completeness of data during this time. The daily sunshine duration is grouped into 6 classes, 0.0, 0.1-3.0, 3.1-6.0, 6.1-9.0, 9.1 - 12.0 and > 12.1 hrs per day. In this work we study the annual distribution of the mean hourly values for each time interval and also the mean monthly values for every meteoorological station separately. Moreover, when we consider as "sunless" those days for which the daily sunshine duration does not exceed 3.0 hrs, we calculate the number of 1, 2, 3 , . . . consecutive sunless days for the cold period of the year, i.e. October, November, December of one year and January, February and March
152
A.A. Bloutsos and P. J. Pennas
Table 1. Geographical Coordinates of the Meteorologieal Stations o f Crete Island and the Mean Daily Sunshine Duration for Every Month and Corresponding Standard Deviation for the Period 1971-1980 Khania = 35°30 ' X = 24°02 ' M Jan Feb Mar Apt May Jun Jul Aug Sep Oct Nov Dec (310) (283) (310) (300) (310) (300) (310) (310) (300) (310) (300) (310) 3.41 3.98 5.93 7.49 9.82 11.98 12.19 11.41 9.28 6.02 4.94 3.58 o 3.07 3.32 3.53 3.92 3.26 1.80 1.15 1.22 2.43 3.53 3.14 3.04 4.32 Suda ¢= 35°33 ' X = 24007 ' M 3.61 4.08 6.03 7.63 9.79 11.78 12.12 11.44 9.29 6.35 5.15 3.74 7.65 o 3.12 3.36 3.63 3.94 3.30 1.96 1.12 1.31 2.54 3.50 3.12 3.00 4.77 Tymbakian ~0= 35°04 ' X = 24°45 ' M 4.69 5.24 6.88 7.81 9.74 11.60 12.16 11.63 9.85 7.31 5.92 4.60 8.13 o 2.79 3.26 3.15 3.65 3.10 1.86 0.94 1.54 2.04 2.94 2.77 2.83 3.79 Iraklion ~o= 35020 ' X = 25°11' M 3.68 4.17 6.02 7.47 9.81 11.86 12.10 11.14 9.37 6.22 5.10 3.79 7.58 o 2.94 3.32 3.46 3.90 3.13 1.97 1.15 1.28 2.22 3.40 2.88 2.96 4.17 Ierapetra ~o= 35°00 ' X = 25°45 ' M 5.10 5.65 6.71 7.62 9.95 12.09 12.36 11.90 11.06 7.65 6.68 5.18 8.34 o 2.67 2.99 3.00 3.64 3.06 1.73 0.99 0.78 1.74 2.79 2155 2.84 3.65 Sitia ~0= 35°12 ' X = 26006 ' M 3.81 4.47 6.23 7.26 9.77
11.86 12.03
o 2~90 3.23 3.27 3.92 3.29
1.83 1.04
10.90 9.22 6.42 5.37 3.89 7.62
1.05 2.05 3.13 2.78 2.94 4.02
Annual(365.3)7.52
o f the following year for each meteorological station. For the winter 1970-1971 only the months January, February and March were used while for the winter 1980-1981 only the months October,, November and December were used. We must also note here that for some spells we calculate the number o f consecutive days using either the last days of September or the first days o f April. Finally we give a theoretical distribution for the nearly sunless spells for each meteorological station using a slightly modified Polya method (Arley et al. 1973) because we found that Polya's m e t h o d does not fit well to empirical data while the modified distribution, using the X2-test, fits better at the 95% confidence level. According to Polya's method the probability of having N ( x ) spells of x consecutive days is given by the equation
Px = P x - 1 ( m + ( x - - 1 ) d ) / x (1 + d ) ,
next one, E ni is the total number o f the consecutive days, S i = x - - 1, x = 1 , 2 , 3, . . The number o f the expected occurrence of an event among N trials is given b y the equation
v = P'N.
The modified Polya m e t h o d gives the number o f spells o f consecutive days using the formulas
2 m e -m
N(1) = k N " (l+d) N(2) = X
m+(x--1)d x(l +d)
•(rn + d)
m+d
(1)
a N(1 ) (2)
4(1 + d)
N(x
N(x) =
1),x=2,3,4,...(3)
where k, X are empirical parameter. The fit o f theoretical to empirical data is tested by using the ordinary ×2-test.
where Po = 1 / ( l + d ) m/a, m = ( N - - ~ n i ) / ~ n i is the expected number o f days (mean), 02 = ( ~ n i s 2 / ~ , n i ) - - m 2 is the variance, N is the total number o f days, d = (o 2/m) 1 is a parameter o f the influence Of an event on the
3. Discussion
3.1 General
Table 1 gives the name o f the meteorological station, its geographical coordinates, the mean
Daily and Hourly Change of Sunshine Duration Values on Crete Island daily sunshine duration values for every month with the corresponding standard deviation as well as the number of observations per month, common to every meteorological station, for the period 1971-1980. As we can see from the above table or as has been mentioned by other scientists, e.g. Pennas (1976), Crete island could be divided, regarding the geographical distribution of sunshine duration values, into two individual parts using an axis from West to East along the mountain crest (Fig. 1), i.e. into the north part and the s o u t h one. Thus, on an annual basis the excess in sunshine duration values of the south side o f the island is about 300 hrs/year while on a monthly basis this could rise up to 50 hrs/month. This difference between the two sides of the island is attributed to the more fregment formation of clouds on the slopes of the mountains exposed to the north winds, due to the forced upward motion of the prevalent north winds during the year in this area (WMO ! 964). But differences in sunshine duration values are also apparent between the meteorological stations of the same zone during the cold months, in that the values for the meteorological stations located on the east edge of eacl~ zone exceed those located on the west side of the same zone. The above differences during the cold period are to be attributed, in our opinion, more to the peculiarities of the atmospheric circulation above the studied area, Machairas (1983), Karoulias (1975) and less to the ground relief. In fact, during the cold period of the year, cyclones of the Mediterranean Sea traverse this area with a general direction from SW to NE and as a result they influence first and more strongly the west side o f thee island. All t h e differences mentioned above are statistically tested at the 95% confidence level and found to be significant.
153
From Fig, 2 we can see that t h e sunshine duration exceeds 30 hrs/month for every hour interval from 08:00 to 16:00 L T (06:00 to 14:00 GMT) for July and A u g u s t . Low sunshine duration values which are apparent during the winter months, especially in the morning and late afternoon hourly time intervals are mainly attributed to the short duration of the day. For the winter months, the sunshine :duration for each hour interval from 08:00 to 16:00 LT has such values that often exceed 10 hrs/month. The isoline of 25 hrs/month for each hour of the day, which correspond to a value of more than 80% of the maximum possible sunshine duration, is between 08:00 to 16:00 LT during the months from May to September for all the examined meteorological stations, while for the meteorological stations located on the east side of the island, these isolines are extended till October.
3.3 Daily Values
Grouping the 3653 daily sunshine duration values for every meteorological station for the decade 1971-1980 into six classes according to the limits mentioned at the beginning of this article, and calculating percentages of occurrence, we draw Fig. 3 (a to f). When we study this figure we can see that, if we exclude the completely sunless days and the days where the sunshine duration is greater than 12 hrs/day, there is not any sufficient difference between the east and west side of the island. But when we study only the completely sunless days (sunshine duration 0.0 hrs/day) we can see that the percentage of their appearance on the meteorological stations of the west side is sometimes twice to three times greater than the percentage for the meteorological stations on the east side of the island.
3.2 Study o f Hourly Values
The mean values of sunshine duration in units of hours per month for each hour of the day and for every month have been calculated from our analytical records and are shown in Fig. 2 (a to f) where for every meteorological station isolines of sunshine duration appear.
3.4 Sunless Spells
Here we are going to examine the nearly sunless spells only during the cold period of the year. As a nearly sunless day we consider a day where the sunshine duration does not exceed 3.0 hrs/ day. The total number of such days for any meteorological station are about 25% of the
154
A.A. Bloutsos and P. J. Pennas: Daily and Hourly Change of Sunshine Duration Values on crete Island
a hrsI . . . . . . . . . . . .
b hrs 1
. . . . . . . . . . .
,, -°\\\\,
o..-
,.! -.:o.X.\-x\
k._)
;/'/o~
hrs/
;
:
:
:
,
,
,
,
,
,
,
i
I0,
o
o
I
e d
~ ~, i
,~ J
J
~
~
8
.
~
J / ~ i
~ ,I -I k ~"6
N~
hr: . . . . . . . . . . . .
h!s f
. . . . . .
Fig. 2. (a-f) Mean monthly sunshine duration (hrs) for each meteorological station
a
1 F F M
b
i
IJ:,~,>~,:~:,:~,~:,;.~l
,;;~, :.'.~,,~7~b~:!'~ !~:7'/~.~"::i ~' ;3 ~' ~.7~:-,,7 ~' "
I: i/Af/llllllltlll.*l~ .
KHANIA
]
i
om ~ ~ ~ ' ; , ) ~ 7 ~ j
.
'!~i:!'~':~.~x,':S'.'-,",~'~Ti, ~:[~./~"); "~'~:,~ ~:~;"."V~
(':.,t
.~:,'-.~,,~'. ,;177
~..;~ .,t:.' D:A~:. ,.t .- 4 :..~,~. ~-~,:~,:,:; ~D',';u
o
mm~)iii!,)){!i ii))))ii~: ~ ; ! 1 :7:~{i ~ ~!;:
oC
io
~
loo;/o
d
ii
~i !!
~ ,~
" i!.
,t~:
x:( '~,-'. ~:.>.:: 7 "c~ )
o
0
~ii:
20 60
/
100% 0 20 60 100%
e
f
l)))iiiii!)i!)iiii)))i))))))i~!l
~! i "~ ~ ~ ~L~;~~,~ ' ~t~+"h.
I ~,~'. 7 "i' "> .")"c2f, r 7Y. L~'~ 1 ~. ;:'D :, D:;:~,., ;:,~, x,, A,~,..I
W. ,~ ~.~
"
I
TM A I N I B KO ]
~ ~{!ii~'7)i;~;l
B, ;, ;,~.s ,v, P,'~-'~'~ ~v :, v :,Z'l
IRPT7 E A ER A
IJ "..r/il/I/I/Ij[fT,.,,G, ~3~+i,~y<')/~'.:3',.::.'!-,~::L~,y ..~ ~ 7' 7.~% ;tt~L,7..i'7 / ~ ..;7:.' :~i
° 17
I ~ / ~
;iil!:~71!}:i!'iTi;~i:i!;!i~il
°
o
m
0,0 0.I-30
NNN
m
).I-60 6.1-9.0 g.l-12J) ~-1211
Fig. 3. (a-f) Percentage of occasions of daily sunshine totals for each meteorological station
156
A.A. Bloutsos and P. J. Pennas
Table 2. Observed and Expected Spells of Consecutive Actually Sunless Days (sunshine duration ~< 3.0 hr/day)as Well as Coefficients and Chi-Square Value for Each Meteorological Station 1 Khania Suda Iraklion Sitia obs exp obs exp obs exp obs exp 2 3 4 5 6 9 8 6 8 7 4 6 4 6 4 5 0 3 1 1 8 9 10 11 12 13 14 15 16 17 2 3 0 4 3 2 2 5 1 4 3 2 2 0 .1 3 2 1 5 2 1 2 1 1 1 1 0 0 2 1 2 1 0 1 0 1 0 1 0 1 1 1 1 1 0 0 0 i 0 0 0 0 0 0 1 0 0 0 1 0 0 0 0 0 1 0 0 .0 1 0 K X Chi-square value 16.263 12.011
150 53 29 14 14 130 47 29 19 12 150 53 29 12 12 140 50 30 19 13
1.5 0.333 1.5 0.333
159 60 39 8 13 10 182 63 36 21 13 8 166 50 30 15 15 135 46 25 14 8 179 65 20 4 198 51 22 10 166 51 12 134 70 17 4 4 4 4 1 1 5 5 0 2
1.5 0 . 3 3 3 32.076 1 0.33 1 0.333 25.024 14.745 14.308
Tymbakion obs exp Ierapetra obs exp
0.5 1
total o f e x a m i n e d days o f the cold period of the decade 1971-1980. The number o f nearly sunless days f o r every meteorological station with 1, 2, 3 etc. consecutive days are given in Table 2. As we can see from this table, the meteorological stations on the south side of Crete island (Tymbakion and Ierapetra) do not have more than one spell o f 9 consecutive nearly sunless days while the meteorological stations on the n o r t h side have at least one spell with 10 or more consecutive days, up to one spell with l7 consecutive days. Another difference between the east and west side o f the island is also apparent in the m a x i m u m number of consecutive sunless days where longer and more spells are observed on the west side. If we omit the spells with more than 3 consecutive sunless days then all the meteorological stations have an almost equal number of spells and there is no difference either between any o f the meteorological stations. We believe t h a t the uniform distribution of the spells with one or two consecutive days is to be attributed to a factor common to all the meteorological stations and this factor we consider to be the influence of moving cyclones which over the examined area have a direction from SW to NE.
On the other hand, the difference in the number o f spells with more than 3 nearly sunless days can be attributed, we believe, to a combination of the ground relief and the prevalence o f n o r t h winds. Using the eqs. (1), (2) and (3) with appropriate k and X coefficient values, given also in Table 2, o f the modified Polya method, the theoretical expected numbers o f spells for meteorological stations are given again in Table 2, while the goodness o f fit is tested using ×2-test at the 95% CL. The results of the above test show that the fit should be considered satisfactory for all the meteorological stations except Iraklion.
4. Conclusions Concluding at this point what we have previously analytically discussed, we could say that to study the sunshine duration, Crete island should be divided into two parts: the n o r t h side which is characterized by a shorter sunshine duration on a daily, m o n t h l y or annual basis and also with longer nearly sunless spells as compared to the south side which is characterized by longer sunshine duration, fewer and shorter spells with sunless days.
Daily and Hourly Change of Sunshine Duration Values on Crete Island There could be a second division o f the island in to the east and west if we consider the m o n t h l y sunshine duration during the cold -period ( O c t o b e r - M a r c h ) where the east part o f the island exceeds that o f the west. This difference between the two parts cannot be observed either during the warm period or in an annual comparison. The primary reason for the difference o b served is the ground relief combined w i t h the prevailing upslope. North winds, over the northern part cause the formation o f clouds while t h e same winds, having downslope characteristics over the southern part help to dissolve clouds.
157
Acknowledgement
The authors wish to thank Dr. Ch. Sahsamanoglou for his very valuable suggestions on this work.
Flocas AA (1980) Estimation and prediction of global solar radiation over Greece. Solar Energy 2 4 : 6 3 - 7 0 Helmes L, Jaenicke R (1984) Experimental verification of the determination of atmospheric turbidity for sunshine recorders. J Clim Appl Met 23:1350-1353 Karapiperis L, Katsoulis B, Papachristopoulos C (1974) Contribution to the study of sunshine duration in Greece. Mere NOA Set II, No 3 8 Karoulias A (1975) The Saharian depressions. Ph D thesis, Univ of Thessaloniki, Greece, in Greek Macris G (1976) On the distribution of solar energy in Greece. Memoirs of the National Observatory of Athens, Ser II. Meteorology 43 Machairas P (1983) Les types des temps d~pressionnaires perturb~s au-dessus de lamer Eg6e. Riv Met Aeronautica XLIII: 13-22 Meteorological Office (1964) Weather in the Mediterranean, vol I, 2nd ed, HMSO, pp 7,119 Pennas P (1976) Sunshine duration in Crete. Sci Ann Fac Phys and Math, Univ of Thessaloniki, 16, 357 WMO (1971) Guide to meteorological instruments and observing practices, no 8 TP3, pp 32-35
References
Arley R, Grisollent H, Guilmet B (1973)Climatologie. M6thodes et pratiques, pp 136-138, 179
Authors' address: Dr. A, Bloutsos and Dr. P. Pennas, Aristotelian University of Thessaloniki, Department of Geology, Sect. Meteorology-Climatology, Thessaloniki 54006, Greece.
