- Email: [email protected]
Geo-spatial distribution of cloud cover and influence of cloud induced attenuation and noise temperature on satellite signal propagation over Nigeria
Accepted Manuscript Geo-spatial distribution of cloud cover and influence of cloud induced attenuation and noise temperature on satellite signal propagation over Nigeria Joseph Sunday Ojo PII: DOI: Reference:
S0273-1177(17)30179-5 http://dx.doi.org/10.1016/j.asr.2017.03.006 JASR 13140
To appear in:
Advances in Space Research
Received Date: Revised Date: Accepted Date:
13 September 2016 27 February 2017 8 March 2017
Please cite this article as: Ojo, J.S., Geo-spatial distribution of cloud cover and influence of cloud induced attenuation and noise temperature on satellite signal propagation over Nigeria, Advances in Space Research (2017), doi: http:// dx.doi.org/10.1016/j.asr.2017.03.006
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
GEO-SPATIAL DISTRIBUTION OF CLOUD COVER AND INFLUENCE OF CLOUD INDUCED ATTENUATION AND NOISE TEMPERATURE ON SATELLITE SIGNAL PROPAGATION OVER NIGERIA Joseph Sunday OJO Department of Physics, Federal University of Technology, Akure, P.M. B 704, Akure, Nigeria. Email: [email protected] or [email protected]
Abstract The study of the influence of cloud cover on satellite propagation links is becoming more demanding due to the requirement of larger bandwidth for different satellite applications. Cloud attenuation is one of the major factors to consider for optimum performance of Ka/V and other higher frequency bands. In this paper, the geo-spatial distribution of cloud coverage over some chosen stations in Nigeria has been considered. The substantial scale spatial dispersion of cloud cover based on synoptic meteorological data and the possible impact on satellite communication links at higher frequency bands was also investigated. The investigation was based on 5 years (2008 – 2012) achieved cloud cover data collected by the Nigerian Meteorological Agency (NIMET) Federal Ministry of Aviation, Oshodi Lagos over four synoptic hours of the day covering day and night. The performances of satellite signals as they traverse through the cloud and cloud noise temperature at different seasons and over different hours of days at Ku/W-bands frequency are also examined. The overall result shows that the additional total atmospheric noise temperature due to the clear air effect and the noise temperature from the cloud reduces the signal-to-noise ratio of the satellite receiver systems, leading to more signal loss and if not adequately taken care of may lead to significant outage. The present results will be useful for Earth-space link budgeting, especially for the proposed multi-sensors communication satellite systems in Nigeria. Keywords: Geo-spatial; Cloud cover; Cloud Attenuation; Cloud noise temperature; Satellite applications; Ku/W-bands; Nigeria 1. Introduction
The technological world is indeed yearning for more transformation on a daily basis and therefore calling for more bandwidth for services that employ radio and satellite telecommunication systems. Most of the new technological devices operating at higher frequency bands enjoy the benefit of higher data transmission rate, smaller antenna size, higher throughput and larger band spectrum to mention but few (Hall, 1980. Ippolito, 2008). However, since the migration from C to higher frequency bands, satellite and terrestrial signal impairment have been the major issues to be solved by the system designers. Studies revealed that satellite communication systems operating at frequency > 10 GHz are more susceptible to rain fade while cloud attenuation has been identified to be one of the major factors that can reduce the usage of Ku/V or higher frequency bands (Allnut and Rogers, 1989, Dissanayake et al., 2001, Sarkah et al., 2005, Das et al., 2013, Ali et al., 2014). Although, rain-induced attenuation plays a major role as far as signal degradation is concerned at these frequency ranges, the effect of cloud attenuation is relatively smaller at Ku and Ka frequency bands. However, at these frequency bands, Earth to satellite links are bound to experience some level of losses due to cloud at any period of the day; depending on the season and the climate of the region (Bouchard, 2008, Ali et al., 2013). Cloud attenuation is even more pronounced at other higher frequency bands, especially those operating at low availability satellite links. This is as a 1
result of the small terminals associated with these bands and the higher probability of occurrence of cloud when compared to rain attenuation. Cloud is known to comprise of tiny water droplets condensing out of the air to form liquid water. In the tropical region, Nigeria inclusive, the occurrence of cloud cover is more prevalent in the year due to the frequency of rainfall in this region when compared to the temperate region. For example, Nigeria climate comprises of about 70% as a yearly average of occurrence of rainy months (some part may sometime experience continuous rainfall for weeks) and thereby liable to more period of cloud cover and associated cloud attenuation (Slobin, 1982, Omotosho and Babatunde. 2010, Ali et al., 2014). Hence, it is paramount to examine the influence of cloud cover and the temperature within the cloud for accurate measurement of cloud-induced attenuation in Nigeria, especially for the design of low fade margin Earth-satellite communication links operating at Ku/V-bands and other higher frequencies. In spite of the attention that the effect of cloud cover has drawn, especially on Earth-satellite communication links at higher frequencies, the subject is just receiving attention in Nigeria. Other applications of cloud cover have always been directed to solar radiation (Ogunjobi et al., 2002, Augustine and Nnabuchi, 2009, Falayi and Rabiu, 2011, Falayi, 2013). For radio and satellite application, for example, Omotosho and Babatunde (2010) examined the impact of cloud on the fixed satellite communication link over some selected stations in Nigeria using satellite data. However, the stations were only categorized on regional basis and not on climatic basis. Also, Omotosho et al. (2011) reported on the effect of cloud cover and cloud attenuation at Ka/V-band frequencies for satellite systems application in tropical wet climate, however the result obtained was too generalized and cannot be specifically applied to geo-climatic region in Nigeria. The seasonal influence, a very strong factor to be considered, especially when determining the level of signal degradation that can be encountered during the worst months was not considered and the effect of noise within the cloud was also not examined. In addition, the attenuation due to cloud obtained was limited to 50 GHz frequency. The significance of the present paper is that it is the first statistical approach that will address the aforementioned shortcomings among others, using a more recent low cloud data obtained from the Nigerian meteorological stations. 2. Data details and Analysis
Daily cloud cover data obtained by the Nigerian Meteorological Agency (NIMET) Federal Ministry of Aviation, Oshodi Lagos, Nigeria from January 2008 to December 2012, have been used in the present study. Micropulse Lidar like conventional radar was used in each of the sites. The instrument comprises of a modified 8-inch telescope with bursts from a laser into the sky directly above the instrument. As the pulses emanating from the laser light travel through the sky, they may come across some tiny water droplets which include a cloud or aerosol particles. The scattered pulses reflect away from the particles and some are redirected back to the instrument below. Based on the time it takes the scattered and reflected pulses to return to the surface, the height of a cloud base can be accurately determined. Usually, the data are collected twice a day, during the day (0.00 hrs and 06.00 hrs) and during the night (12.00 hrs and 18.00 hrs) local time (LT). The low cloud amount is measured in eighths having a range from 0 (no cloud) to 8 (complete cover of sky). The synoptic data values recorded at NIMET stations were available for about 95% of the time. The missing or errors in the data were less than 5%. The cloud data obtained from NIMET have advantage of the availability of a larger number of stations in Nigeria over a large period rather than extrapolating data for a specific station using satellite data. Ten meteorological stations covering some selected cities in Nigeria were used with each of the sites located widely across Nigeria. The sites are considered as representative of different climatic regions (Rain forest belt-RFB, Sudan savannah-SS, Wooden belt-WB, Guinea Sudan-GS and Sahel-S) in Nigeria. Table 1 presents the 2
characteristics of the study sites. Other atmospheric parameters measured alongside the cloud cover are pressure, temperature and dew point temperature at different heights. In order to examine signal attenuation due to cloud, knowledge about the statistics of these clouds for an extensive period of time is required. Cloud is known to exist in different forms such as the, Cumulus (Cu), Stratocumulus (Sc), Cumulonimbus (Cb), Nimbostratus (Ns), Stratus (St), Altostratus (As) and so on. Each of this cloud type has their peculiar characteristics in terms of shape, effective radius, moisture content and liquid water contents among others. Detailed description is not discussed in this report due to paucity of space, but they are readily available in the literature (Slobin, 1982, Sarkah and Kumar, 2007, Garcia et al., 2008, King et al., 2013). Table 1: Characteristics of the study sites Climatic Region
Stations
Coordinates (Long oN, Lat oE)
Altitude (m)
Rain Forest Belt (RFB)
Calabar
4.9757, 8.3417
32
Uyo
5.0377, 7.9128
50
Benin
6.3350, 5.6037
88
Asaba
6.2059, 6.6959
52
Ibadan
7.3775, 3.9470
230
Makurdi
7.7322, 8.5391
114
Yola
9.2035, 12.4954
599
Minna
9.5836, 6.5463
350
Damaturu
11.7470, 11.9662
456
Sokoto
13.0059, 5.2476
356
Wooded Belt (WB)
Guinea Sudan (GS)
Sudan-Sahel (SS)
Sahel (S)
For the spatial distribution of cloud cover, the procedure stated in the work of Garcia et al. (2008) was adopted. Cloud cover is usually expressed in terms of octas, a unit that represents one-eighth of the sky vault. The sum of solid angles Ω, on the sky dome as depicted by the cloud cover varies from 0 (in case of clear sky) to Ωo (in case of overcast sky). Cloud cover amount (or cloud fraction) is then the ratio Ω/ Ωo, which may also vary between 0 and 1, measured in octas (Garcia et al., 2008). The samples of the time series derived from the synoptic meteorological data have been estimated based on two different approaches that correspond to the following modes: Mode 1: Each sample, Xn, of the time series x is defined as the amount of the sky vault covered by a specific type of cloud at the period of measurements, represented as: Xn =
On 8
(1)
where On = 0,1,2,….,8 is the number of octas in the nth data record. The fraction of the cloud cover obtained using equation (1) can be applied in meteorology. Mode 2: For a specific cloud type, the binary time series x is defined as:
X
n
=
O n ≥Ot 1 if 0 if O < O n t
{
(2)
3
where On is the number of octas in the nth data record (or time interval) and Ot is a threshold number of octas, Ot = 1,2,….8. Equation (2) provides information about the threshold level of the number of octas. For the spatial distribution of cloud cover, we have adopted the occurrence of cloud in pairs of sites at each of the climatic regions in Nigeria. This could be achieved by using the statistical metric measure like correlated coefficient, statistical dependence index and the joint probability of simultaneous occurrence of cloud of a specific type over each of the regions (Garcia et al., 2008). As defined in the work of Barbaliscia et al. (1992) and Garcia et al. (2008), if there exists two sites Q and R in the same region, the correlated coefficient can be estimated as: cov ( x , y ) ρ= (3) σ x ,σ y where x and y are the time series earlier defined in equations (1) and (2) for sites Q and R, respectively. The statistical dependence indexχQR, is also expressed as:
χ QR =
PQR
(4)
(P .P ) Q
R
where PQ and PR are the individual probabilities of cloud cover of a specific type at the sites Q and R, respectively. The joint probability of the simultaneous occurrence of cloud covers of a specific type, in the nth record, PQR is given as:
PQR =
1 n ∑ ( X n .Yn ) N n =1
(5)
where Xn and Yn are same as defined in equations (1) and (2). The systematic formulation of attenuation due to cloud at microwave and millimeter wave frequency follows nearly the same method as the method used in predicting attenuation due to rain. The significant difference between the two atmospheric parameters (rain and clouds) is that clouds are made up of suspended mist of water drops with very small diameters less than 1µm while rain is the liquid water in the form of droplets that have condensed from atmospheric water vapor and become heavy enough to fall under gravity (Hall, 1980, Ali et al., 2014). The degradation or impairment due to the suspended water droplets contained in atmospheric clouds can be estimated with some level of accuracy, using the model developed by Liebe (1989) for fog and cloud attenuation at frequencies up to 1000 GHz based on the Rayleigh scattering of electromagnetic wave, which is associated with a double-Debye model for the dielectric permittivity of water. The model has generally been adopted by ITU-R Recommendation ITU-R P.840-5 (2012). This same procedure is adopted in this present study as expressed step by step below: The expression for the specific attenuation within a cloud or fog, γc can be written as:
γ c = Kl M
(6)
M which is the liquid water density in the cloud (gm-3) is obtained from the radiosonde profiles at different NIMET stations. For the purpose of this work, M is estimated using the Salonen’s model from radiosonde data at each of the selected stations. Cloud’s formation is based on the knowledge of occurrence of high relative humidity whereby the radiosonde data show the evidence of LWC based on whether the relative humidity has gone beyond its critical value. According to Karsten’s 4
model, the formation of cloud occurs when the relative humidity exceeds 95% and the phase is determined by the temperature profile (Chakraborty and Animesh, 2012). Therefore, by using the thermodynamic concept, the LWC for each height level can be estimated from the following expression: Cp (Γd − Γs )dz M ad ( h ) = ∫ ρ (z ) (7) Lh where: Mad(h)= liquid water content under adiabatic condition; Lh = latent heat of vaporization given as 80 cal/gm; Cp = specific heat at constant pressure with the value 1.0035 J·g–1·k–1; ρ(z) = air density; Γd = dry adiabatic lapse rate; Γs = moist adiabatic lapse and varies from 4˚C/km to 9.8˚C/km depending on the seasonal variation of temperature. Hence, the adiabatic condition produces a maximum value of the LWC that is given by equation (8) as a result of the circulation of air mass accompanied by precipitation and freezing temperatures.
M = M ad (1.239 − 0.145 ln ∆h ) (8) where ∆h = height above the cloud base. M is then calculated at each pressure level at a particular radiosonde ascent. Integrating the LWC profile over height, the total value of the LWC is obtained at each ascent (Chakraborty and Animesh 2012). Also, the parameter Kl in equation (6) is the specific attenuation coefficient (dBkm-1)/gm-3) and based on Rayleigh scattering can be expressed as:
Kl =
0.819 f (dBkm-1) 2 ε ′′(1 + η )
(9)
where f is the frequency (GHz) , and
η=
2+ε′ ε ′′
(10)
where ε ′ and ε ′′ are the complex dielectric permittivity of water as a function of frequency and temperature. The parameters can be obtained using double-Debye model (Das et al., 2013). To obtain the attenuation due to clouds for a given probability of time, the statistics of the total columnar content of liquid water L (kg/m2) or, equivalently, mm of precipitable water for a given site must be known yielding: A=
LK l sin θ
(dB) for 90o > θ > 5o
(11)
where θ is the elevation angle and Kl is given as in equations (7) - (15) and The total atmospheric noise temperature Ttatm due to clear air effect (addition of Oxygen and water vapour) and the noise temperature from the cloud for different cloud thickness up to 2 km are also estimated for the selected stations during the worst month’s condition using (Slobin, 1982; Sarkar and Kumar, 2007): 1 Ttatm = T f 1 − (12) Ls and Ls = 10 [αtotal ( dB ) ] / 10 (13)
5
where Tf is the physical temperature, Ls is the loss factor, and αtotal is the total attenuation of radio wave due to oxygen, water vapour and cloud. For the purpose of this work, the attenuations of radio wave due to oxygen and water vapour are estimated using ITU-R P. 676-10 (2013) and processed with Matlab 2016a gaspl function for the selected frequencies over each of the study locations. 3. Results and discussion This section discusses in detail results on the morphology of cloud occurrences, spatial distribution of cloud cover, noise temperature within the cloud and their possible effect on Earth-satellite signal links at higher frequencies over the selected sites in Nigeria.
3.1. Cloud cover morphology for the selected sites Cloud occurrence morphology is necessary to understand the extent of degradation on signal traversing through Earth-satellite path and for optimizing power sharing in multi-sensors satellite systems. Clouds are known to consist of liquid water particles having diameters from about 0.1 to about 1.0 gm-3, while a typical radius of water droplets within non-rainy clouds varies between about 1 µm and 30 µm with a maximum droplet density of about 3-6 µm (Ippolito, 2008, Mandeep and Hassan, 2008). A particular cloud type will have a range of water particle sizes, while the thickness of the cloud varies from about 1.5 to 2.5 km depending upon the type of the cloud as reported by Kumar and Sarkar, (2007). As earlier reported in section 2, there exist several types of cloud, among which Cu/Cb are very important in Nigeria due to their characteristics. Cumulus clouds are low-level cloud with flat bottoms and rounded tops, and grow vertically. It is known to contain particles with diameters from about 4 to 15 µm; cumulonimbus clouds have particle diameters from 2 up to 100 µm, where the distinction between cloud particles and suspended rain is not clear. Cumulus cloud types with large vertical growth generally have large liquid water content while cumulonimbus results from when enough atmospheric instability, moisture, and lift are present. That often leads to a strong updraft within the cumulus cloud resulting to a mature and deep cloud. Cloud electrification occurs within the cumulonimbus clouds due to many collisions between charged water droplets, graupel (ice-water mix), and ice crystal particles, resulting in lightning and thunderstorm which is one of the characteristics of tropical Nigerian climate. Based on the local data obtained from NIMET, Table 2 presents the cloud cover for individual sites. The time-series have been calculated for each of the modes earlier described in equations (1) and (2). The time-averaged cloud amount is calculated by summing the samples for each type of cloud and dividing by the total number of synoptic records. The results from Table 2 revealed the climatic peculiarities expected at each site. It is seen that the occurrence of Cu/Cb cloud types is the more frequent in the synoptic records, the corresponding values in the table are relatively close to the values of the “low cloud” group. The larger cloud amounts are observed in southern regions within the rain forest climatic zone (Calabar, 52.6%), whereas sites located in the northern region of Nigeria gave the smaller cloud cover amounts (Damaturu, 14.9%). Ns presence is also more common in the Southern regions often affected by rainy fronts, which agree with the fact that stratiform rains are usually associated with Ns clouds. The present result do not agree with those earlier obtained in Spain (Temperate region) where a higher percentage of cloud cover occurs in the northern regions with St/Sc cloud type being the most frequent in the synoptic records (Gracia et al., 2008). This shows the dynamical nature of cloud cover and how geographical location can dictate the occurrence of cloud cover. A more random distribution is noted in the case of St or Sc clouds. The values in the right-hand side of Table 2, give an indication of how often the threshold “no cloud” has been exceeded. The Southern region has a close proximity to the Sea, (in this case the Atlantic Ocean) which produces more cloud as earlier reported on the work of Garcia et al. (2008). There is over 70% occurrence of cloud 6
cover in the Southern regions. Also, the most global variance in cloud radiative response between Global Cloud Models (GCMs) is due to low clouds with about 50% due to the cumulus regime and 20% due to the regime characterized by clouds undergoing transition from cumulus to cumulonimbus. Table 2. Cloud cover for individual sites Mode 1
Mode 2
Climatic Region
Stations
Low (%)
Cu/Cb (%)
St/Sc (%)
Ns (%)
Low (%)
Cu/Cb (%)
St/Sc (%)
Ns (%)
RFB
Calabar
52.1
44.1
5.0
7.9
73.0
49.5
14.8
9.4
Uyo
51.6
27.8
3.2
7.7
70.2
56.2
13.5
5.8
Benin
41..8
21.8
2.9
5.8
41.0
30.0
6.6
3.3
Asaba
42.2
33.3
4.0
6.3
50.8
36.8
7.2
4.9
Ibadan
41.8
31.9
3.9
6.0
50.1
33.4
7.1
3.6
Makurdi
38.5
7.0
1.3
2.1
46.6
39.9
5.5
2.8
Yola
33.6
26.6
3.2
5.0
42.5
37.7
8.7
8.2
Minna
32.1
19.8
2.2
4.0
38.6
30.4
6.5
4.9
Damaturu
14.9
10.2
5.6
5.2
19.3
10.7
8.5
1.1
Sokoto
18.9
12.0
1.5
3.6
25.8
19.3
3.4
1.6
B
GS
SS
S
The cloud cover observations over the four times of the day comprising the daytime and the nighttime (06:00, 12:00, 18:00 and 00:00 hrs LT) are also characterized and results are presented in Figs. 1 (a-b) for RFB sites, Figs 2 (a-b) for WB sites, Figs 3 (a-b) for GS sites, Figs 4 (a-b) for SS sites and Figs. 5 (a-b) for S sites during daytime and the nighttime respectively. For ease of result analysis, the sites are paired alongside the climatic region as earlier categorized in section 2. For example Figs. 1 (a-b) present the monthly distribution of cloud cover over RFB region during daytime and nighttime respectively. In Fig. 1(a) for example, the result shows that over the RFB region, daytime low cloud is significant in the months of June, July, August and September at 6 hrs for 19, 21, 24 and 22 days and at 12 hrs LT for 18, 19, 21 and 19 days in Calabar while daytime low cloud is found to occur in 15, 17, 22 and 18 days and 14, 14, 16 and 18 days in Uyo during the respective hours and same months. It is however noticed that, irrespective of the months, daytime low cloud covers occurred more frequently in the 6 hrs LT when compared with the 12 hrs LT. The result for the nighttime low cloud as presented in Fig. 1 (b) also shows that, over the RFB coastline, the sky is partially covered with clouds in the months of April, June, July, August, September and October for 16, 19, 19, 21, 15 and 16 days in Calabar and 14, 16, 15, 19, 16 and 14 days in Uyo during midnight (12 hrs LT). Further results show that at the 18 hrs LT, the sky is covered with clouds for 11, 9, 10 and 11 days in the months of April, August, September and 7
October respectively and 9, 5 and 6 days in the months of April, August and October respectively in Calabar. It is evident here that the effect of ocean breeze is more influenced in Uyo in the nighttime when compared with Calabar. It could also be seen that Figs 1(b) and 3 (b) have a variation compared to some extent to Figs 1(a) and 3(a), this is may be due some similar climatic features between the two climatic regions (Rain forest belt and Guinea Sudan). For example, both regions have a moderate occurrence of the average rainfall amount due to cloud-cover when compared with other climatic regions considered in this work. Other climatic regions follow the same trend, although with different numbers of days the sky is covered with clouds and probably different months of the year as presented in Figs. 2 – 5. Uyo (6 hrs LT) Average (6 hrs LT) Calabar (12 hrs LT)
25 NUMBER OF DAYS
NUMBER OF DAYS
25
Calabar (6 hrs LT) Uyo (12 hrs LT) Average (12 hrs LT)
20 15 10 5
Uyo (18 hrs LT) Average (18 hrs LT) Calabar (00 hrs LT)
Calabar (18 hrs LT) Uyo (00 hrs LT) Average (00 hrs LT)
20 15 10 5
0
0 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec MONTH
MONTH
Fig. 1: Monthly distribution of cloud coverage over RFB region during (a) daytime and (b) nighttime Benin (6 hrs LT) Average (6 hrs LT) Asaba (12 hrs LT)
25 NUMBER OF DAYS
NUMBER OF DAYS
25
Asaba (6 hrs LT) Benin (12 hrs LT) Average (12 hrs LT)
20 15 10 5
Benin (18 hrs LT) Average (18 hrs LT) Asaba (00 hrs LT)
Asaba (18 hrs LT) Benin (00 hrs LT) Average (00 hrs LT)
20 15 10 5 0
0 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec MONTH
MONTH
Fig. 2: Monthly distribution of cloud coverage over WB region during (a) daytime and (b) nighttime Ibadan (6 hrs LT) Average (6 hrs LT) Makurdi (12 hrs LT)
20
25
Ibadan (18 hrs LT) Average (18 hrs LT) Makurdi (00 hrs LT)
Makurdi (18 hrs LT) Ibadan (00 hrs LT) Average (00 hrs LT)
NUMBER OF DAYS
20
15
15
10
10
5
5 0
Dec
Nov
Oct
Sep
Aug
Jul
Jun
May
Apr
Feb
Mar
0 Jan
NUMBER OF DAYS
25
Makurdi (6 hrs LT) Ibadan (12 hrs LT) Average (12 hrs LT)
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec MONTH
MONTH
Fig. 3: Monthly distribution of cloud coverage over GS region during (a) daytime and (b) nighttime. 8
Yola (6 hrs LT) Minna (12 hrs LT) Average (12 hrs LT)
NUMBER OF DAYS
NUMBER OF DAYS
18 16 14 12 10 8 6 4 2 0
Minna (6 hrs LT) Average (6 hrs LT) Yola (12 hrs LT)
18 16 14 12 10 8 6 4 2 0
Minna (18 hrs LT) Average (18 hrs LT) Yola (00 hrs LT)
Yola (18 hrs LT) Minna (00 hrs LT) Average (00 hrs LT)
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
MONTH
MONTH
18 16 14 12 10 8 6 4 2 0
Damaturu (6 hrs LT) Average (6 hrs LT) Sokoto (12 hrs LT)
Sokoto (6 hrs LT) Damaturu (12 hrs LT) Average (12 hrs LT)
NUMBER OF DAYS
NUMBER OF DAYS
Fig. 4: Monthly distribution of cloud coverage over SS region during (a) daytime and (b) nighttime
18 16 14 12 10 8 6 4 2 0
Damaturu (18 hrs LT) Average (18 hrs LT) Sokoto (00 hrs LT)
Sokoto (18 hrs LT) Damaturu (00 hrs LT) Average (00 hrs LT)
MONTH
MONTH
Fig. 5: Monthly distribution of cloud coverage over S region dur during ing (a) daytime and (b) nighttime
3.2.Statistical evaluation of the spatial distribution of cloud cover To understand the spatial distribution and occurrence of cloud cover, we have adopted the sites pairing method using the correlation coefficient and statistical dependence index based on the climatic region in Nigeria. Results are presented in the form of maps to give distinct information needed for link behavior of cloud cover on each of the sites. The maps are developed using Kriging Interpolation Technique (KIT) by incorporating all the statistical results using equations (3) –(6). KIT was used for the spatial interpolation of correlation coefficient and statistical dependence index cloud cover values based on regular grid. This will assist to generate a highly consistent and predictable inter-site cloud cover variation for any of the geographical location within the region under investigation The contour maps are developed based on the results presented in each of the columns in Table 2. Fig. 6 (a) presents the map for mode 1, for the low cloud group of the cloud cover amount over the regions. If mode 2 is used, a threshold of 1 octa should be considered and the probability of detecting any amount of cloud in any sky region for the low cloud group can also be obtained. The result is plotted on the map represented in Fig. 6(b). Each of the maps has different shape of the contour lines based on the climatic characteristics of each of the geographical region. For example the rainforest zone, which is close to the sea produces relatively large difference in cloud cover between sites with closer proximity. This is evident in the 9
cloud cover values of about 50% in Calabar near the sea shore which reduces below 40% for Asaba site moving upward the sea thereby creating a steep gradient in contour lines. The map derived using mode 2 with a threshold of 1 octa shows an increase in cloud detection in coastal and mountainous areas. This may be often due to cumulus cloud reports, which may be visible in small sky regions in the RFB and WB sites. The cloud cover amount reduces as we moved towards the northern part of the country. (%)
12
(%)
12
80
11
11
52
75 48
10
10
70
9
40 36
8
L A T IT U D E (D E G )
L A T IT U D E (D E G )
44
65
9
60 55
8
50
32 7
7
45
28 6
24
40
6
35 20
30
5
5
16 4 5
6
7
8
9
10
11
12
13
12
LONGITUDE (DEG)
25 4 5
6
7
8
9
10
11
12
13
20
LONGITUDE (DEG)
(a)
(b)
Fig. 6: Map of cloud cover value for the low cloud group using (a) mode 1 and (b) mode 2. Figures 7 (a) and (b) also show the map of the correlation coefficients and the dependence index respectively for the various stations. The contour map is plotted for the sites by taking into account the values relative to the inter-site under investigation. The value for these parameters diminishes with distance as we move from the southern coastline to the northern plain sites; the shape of the contour lines is strongly influenced by climatic and geographical factors of the selected sites. This is evident in the contour lines, because if only the influence of distance is considered, the contour will just be circling lines. For example the correlation coefficient maps depicted in Fig. 7(a) shows that the contour lines expand along the RFB coastline (Calabar and Uyo) but compressed towards the GS which is more mountainous topographically and enlarging towards the relatively flat northern part of the country (SS and S). The contour lines around the WB region (Asaba and Benin) tend to follow the RFB bearing the narrow coastal strip. These trends are also observed in the statistical dependence index (χ) map as presented in Fig. 7 (b). The implication of the maps is in the application of the geo-spatial distribution analysis in the design of the multi-sensors satellite systems over the study sites and application of the site diversity techniques among others. For example the information from the maps will enable communication system designers to cater for the level of signal degradation in terms of sharing the available power (downlink or uplink) to mitigate the effects of simultaneous bad weather in the given section of the coverage area. As noted in the work of Garcia et al. (2008), if a given site X1 needed additional power due to cloud attenuation, the statistical dependence index values estimated for the pair of sites X1 –X2 will show the probability that additional power is also needed in the link to site X2. Also, information from the maps will assist to know the level of site dependency as well as the degree of relation between sites. For example, if χ is equal to or approaches 1, statistical independence is achieved and both sites can be dealt with independently. However, higher values of
10
χ indicate that both sites are prone to cloud attenuation. The same trends are applicable to the correlation coefficients (ρ) bearing in mind the diverse range of ρ. 12
12
11
11
2
0.95
10
10
1.9
0.9
1.8
0.8 0.75
8
0.7 0.65
7
0.6
LATITUDE (DEG)
LATITUDE (DEG)
0.85 9
9
1.7 1.6 1.5
8
1.4 1.3
7
1.2
0.55 0.5
6
1.1 6
1
0.45 0.4
5
0.9 5
0.8
0.35 0.3
4 5
6
7
8
9
10
11
12
0.7 0.6
4 5
13
6
7
8
9
10
11
12
13
LONGITUDE (DEG)
LONGITUDE (DEG
(a)
(b)
Fig. 7: Statistical map of (a) correlation coefficients and (b) dependence index relatives to the study sites 3.3.Specific attenuation and total attenuation due to cloud coverage
The specific attenuation due to cloud has been estimated for each of the sites at different climatic region based on ITU-R P.840-5 (2012) method, using equations (6) – (10) on the basis of the liquid water content 1 gm-3 and for a cloud temperature of 273 K as input parameters. The estimations are done over frequency ranges from the Ku to W-band following the IEEE Standard 521-2002, namely: 12, 20, 30, 50, 70 and 100 GHz. Results are presented in Table 3 for the specific attenuation on radio wave due to cloud cover in the selected sites across Nigeria. It is seen from Table 3 that the specific attenuation over the RFB coastline for water content ~ 1 gm-3 range between 0.02 and 5.32 dB/km at Ku – W band frequencies (12-100 GHz) for Calabar and between 0.18 – 4.89 dB/km at Uyo. It has earlier been noticed that the variation in the specific attenuation is due to the variation of cloud particle temperature (Sarkar et al., 2006). It is worth mentioning that the specific attenuation due to cloud decreases from the southern coastline region, which is characterized by more occurrence of cloud cover to the plain northern part of the country that is characterized by lesser cloud coverage. This is clearly evident in the result presented in Table 3 for all the stations at frequencies 12 – 100 GHz. Following the procedure highlighted in ITU-R P.840-5 (2012) using equations (6) – (11), the attenuation due to the cloud over the selected sites were estimated. The estimation was based on the total columnar content of liquid water L (kgm-2) as stated in equation (11). The NIGCOMSAT-R1 elevation angle of 42.5o has been used since it is an indigenous satellite for Nigerian communication visible over all the stations. The results of the total attenuation due to cloud at the specified frequency range for different cloud thickness of 1 km, 1.5 km and 2 km are presented in Table 4. The values have been chosen to conform to the low cloud thickness (Cu/Cb) associated with Nigeria climate. Generally speaking, the cloud attenuation observed over the locations show that the values are lower in Ku-band frequency and higher at W-band frequency. It is also evident from the results that cloud-induced attenuation increases as cloud thickness increases. The implication is that uplink frequencies will be grossly attenuated from the Earth-satellite paths even during clear sky conditions. The results, based on the climatic categorization and cloud thickness show that even at the same climatic region, mostly at V and W-band frequencies, the differences in 11
cloud attenuation value can be as high as 3 dB most especially at the RFB coastline even during clear air conditions. Some results earlier obtained by Omotosho and Babatunde (2010) were also corroborated in this study. For example, it was reported in the paper that cloud attenuation is generally higher in the coastal region than in the plain Northern region of Nigeria and that at Vband, the degradation may become a major concern for satellite links especially in the coastline region. The information presented in Table 4 will also assist in estimating atmospheric noise generated by the cloud which serves as additional signal degradation to Earth-satellite links even under clear weather conditions. 1.1. Enhanced noise temperature due to cloud
The total atmospheric noise temperature was also estimated to determine the degree of extra noise due to the noise generated by cloud and the consequences on satellite signal receivers in the region. The results on the atmospheric noise temperature with thickness 1, 1.5 and 2 km using equations (12) and (13) are presented in Figs. 8 (a-e) for RFB, WB, GS, SS and S region respectively. As usual, the sites are paired based on the climatic categorization. For example, in Fig 8 (a) for the RFB region, the total atmospheric noise temperature over Calabar from 1 to 2 km thickness varies between 15.37 and 24.57 K at 12 GHz Ku-band frequency, 63.22 and 87.63 K at 20 GHz K-band frequency, 85.58 and 110.45 K at 30 GHz Ka-band frequency, 199.43 and 286.64 K at 50 GHz Vband frequency, 406.67 and 459.56 K at 70 GHz and 648.34 and 741.52 K at W-band frequency. Uyo which also belongs to RFB region shares similar values of total atmospheric noise temperature from 1 to 2 km thickness and the values vary between 14.35 and 22.30 K at 12 GHz Ku-band frequency, 61.09 and 83.45 K at 20 GHz K-band frequency, 82.34 and 102.35 K at 30 GHz Kaband frequency, 198.34 and 272.36 K at 50 GHz V-band frequency, 404.26 and 457.63 K at 70 GHz and 643.41 and 736.23 K at W-band frequency. The trend continues in other climatic regions, although with different values of total atmospheric noise temperature. The implication of the results is that the additional total atmospheric noise temperature due to the clear air effect and the noise temperature from the cloud reduces the signal to noise ratio (SNR) of the satellite receiver systems, leading to more signal loss and if not adequately taken care of may lead to significant outage. The effect may even be more pronounced for low noise receiving satellite systems.
12
Table 3: Specific attenuation of radio wave due to cloud cover in Nigeria. Station
Calabar
Uyo
Asaba
Benin
Makurdi
Ibadan
Yola
Minna
Damaturu
Sokoto
Frequency (GHz) 12 20 30 50 70 100 12 20 30 50 70 100 12 20 30 50 70 100 12 20 30 50 70 100 12 20 30 50 70 100 12 20 30 50 70 100 12 20 30 50 70 100 12 20 30 50 70 100 12 20 30 50 70 100 12 20 30 50 70 100
Specific attenuation (dB/km) 0.22 0.42 1.34 2.65 4.24 5.35 0.18 0.38 1.11 2.31 4.01 4.89 0.15 0.36 0.89 1.89 3.91 4.04 0.11 0.32 0.81 1.53 3.75 4.02 0.16 0.37 0.83 1.63 3.67 4.03 0.14 0.28 0.82 1.59 3.23 3.89 0.08 0.23 0.67 1.49 3.18 3.83 0.06 0.21 0.58 1.41 3.07 3.76 0.05 0.2 0.49 1.38 3.02 3.29 0.06 0.18 0.47 1.22 3.01 3.19
13
Table 4: Cloud-induced attenuation of radio wave at different cloud thickness over Nigeria Station
Calabar
Uyo
Asaba
Benin
Makurdi
Ibadan
Yola
Minna
Damaturu
Sokoto
Frequency (GHz) 12 20 30 50 70 100 12 20 30 50 70 100 12 20 30 50 70 100 12 20 30 50 70 100 12 20 30 50 70 100 12 20 30 50 70 100 12 20 30 50 70 100 12 20 30 50 70 100 12 20 30 50 70 100 12 20 30 50 70 100
1.0 km 0.43 0.89 1.95 5.26 7.92 10.84 0.32 0.75 1.82 4.89 6.62 9.89 0.21 0.42 1.22 4.01 6.11 9.13 0.18 0.31 1.01 3.82 5.02 8.22 0.2 0.38 1.12 3.97 5.14 8.64 0.13 0.24 1.02 3.11 5.08 8.19 0.1 0.19 0.92 3.06 4.98 8.04 0.11 0.21 1.01 3.08 5.01 8.04 0.09 0.16 0.84 2.76 4.23 7.23 0.06 0.14 0.46 2.14 4.01 6.24
14
1.5 km 0.67 0.98 2.05 5.74 8.12 10.97 0.42 0.85 2.01 4.99 6.87 9.96 0.42 0.65 1.34 4.45 6.82 9.23 0.22 0.53 1.21 3.98 5.65 8.75 0.26 0.45 1.34 4.03 5.58 8.96 0.31 0.42 1.39 3.42 5.78 8.89 0.22 0.36 1.26 3.22 5.08 8.21 0.19 0.31 1.18 3.17 5.14 8.35 0.12 0.28 0.98 2.97 4.82 7.68 0.08 0.21 0.78 2.35 4.32 6.85
2.0 km 0.74 1.23 2.24 5.89 10.76 14.83 0.65 0.98 2.73 5.28 9.24 12.26 0.48 0.76 1.42 4.56 7.98 11.25 0.32 0.65 1.54 4.23 6.87 10.32 0.35 0.72 2.01 4.36 7.22 10.87 0.38 0.54 1.89 3.98 6.22 9.89 0.28 0.45 1.67 3.65 6.12 9.57 0.21 0.42 1.43 3.32 5.89 9.14 0.14 0.32 1.14 3.01 5.22 8.45 0.09 0.28 0.96 2.89 5.01 7.56
800
(a)
700 600 500 400 300 200 100 0
Asaba-1 km cloud thickness Asaba-1.5 km cloud thickess Asaba-2.0 km cloud thickness Benin- 1km cloud thickness Benin-1.5 km cloud thickess
600 500 400 300 200 100 0
0
50 100 FREQUENCY (GHz)
150
0
800
800
(c)
700
Mak-1 km cloud thickness Mak-1.5 km cloud thickess Mak-2.0 km cloud thickness Iba- 1km cloud thickness Iba-1.5 km cloud thickess
600 500 400 300 200 100 0
50 100 FREQUENCY (GHz)
Yola-1 km cloud thickness Yola-1.5 km cloud thickess Yola-2.0 km cloud thickness Minna- 1km cloud thickness Minna-1.5 km cloud thickess
600 500 400 300 200 100 0
50 100 FREQUENCY (GHz)
ATMOSPHERIC NOISE TEMPERATURE (K)
0
150
800 700 600 500 400 300 200 100 0
150
(d)
700 ATMOSPHERIC NOISE TEMPERATURE (K)
ATMOSPHERIC NOISE TEMPERATURE (K)
(b)
700
Cal-1 km cloud thickness Cal-1.5 km cloud thickess Cal-2.0 km cloud thickness Uyo- 1km cloud thickness Uyo-1.5 km cloud thickess
ATMOSPHERIC NOISE TEMPERATURE (K)
ATMOSPHERIC NOISE TEMPERATURE (K)
800
0
50 100 FREQUENCY (GHz)
150
(e) Dam-1 km cloud thickness Dam-1.5 km cloud thickess Dam-2.0 km cloud thickness Sok- 1km cloud thickness Sok-1.5 km cloud thickess
0
50 100 FREQUENCY (GHz)
150
Fig. 8: Total atmospheric noise temperature with thickness 1, 1.5 and 2 km for different frequencies over (a) RFB (b) WB (C) GS (d) SS and (e) S. Conclusion In this paper, 5 years (2008 – 2012) recent archived cloud cover data collected by the Nigerian Meteorological Agency over four synoptic hours of the day covering day and night have been used to evaluate the spatial distribution of cloud cover and the effect of cloud attenuation and cloud noise temperature on the performance of Earth-satellite links across the Geo-climatic region of Nigeria. The results on spatial distribution of cloud cover shows that the occurrence of Cu/Cb cloud type is the most frequent in the synoptic records, with the larger cloud amounts observed in southern regions with rain forest climatic zone while the sites located in the northern region of Nigeria gave 15
the smaller cloud cover amounts. The results on cloud cover observations over the four times of the day comprising the daytime and the nighttime also show that irrespective of the months, daytime low cloud covers occurred more frequently in the 6 hrs LT when compared with the 12 hrs LT while nighttime low cloud covers occurred more frequently in the 00 hrs LT when compared with the 18 hrs LT. Contour maps generated for the correlation coefficients and the relative dependence index will be useful in the design of the multi-sensor satellite systems over the study sites and application of the site diversity techniques among others. The results presented for the specific attenuation due to cloud revealed decreases in value from the southern coastline region, which is characterized by more occurrence of cloud cover to the plain northern part of the country that is characterized by lesser cloud coverage. Also, results on total attenuation shows that even in the same climatic region, mostly at V and W-bands frequency, the difference of cloud attenuation can be as high as 3 dB most especially at the RFB coastline even during clear air conditions. When the coastline is compared with plain northern sites, the difference can be as high as 7 dB at W-band frequency and 2 km cloud thickness. Results on additional total atmospheric noise temperature due to clear air effect and the noise temperature of the cloud lower the signal-to-noise-ratio of the satellite receiver systems, leading to more signal loss and if not adequately taken care of may lead to significant outage. Acknowledgements
The author wishes to acknowledge the Nigerian Meteorological Agency for the data used in this work and more especially for the International Centre for Theoretical Physics (ICTP), Trieste Italy for the latest software and access to some of the materials used. The effort of Professor O.S Ajayi to proof read the manuscript is also acknowledged. References
Ali Mohammed Al-Saegh, A. Sali, J. S. Mandeep, Alyani Ismail, Abdulmajeed H.J. Al-Jumaily and Chandima Gomes, Atmospheric Propagation Model for Satellite Communications, http://dx.doi.org/10.5772/58238, INTECH, 249-273, 2014. Ali Mohammed Al-Saegha, Aduwati Salia, J.S. Mandeep and Alyani Ismaila, Extracted atmospheric impairments on earth-sky signal quality in tropical regions at Ku-band, Journal of Atmospheric and Solar-Terrestrial Physics, 104, 96-105, 2013 Allnutt J. E, and D. V. Rogers, Low-fade-margin systems: propagation considerations and implementation approaches. IEE Conf. Publ. 301, 6–9, 1989. Augustine C and Nabuchi M.N, Correlation of cloudness index for four selected cities in Nigeria, The pacific Journal of science and Tech, 10(2), 568-573, 2009. Barbaliscia F, Ravaioli G, Paraboni A, Characteristics of the spatial statistical dependence of rainfall rates over large areas. IEEE Transactions on Antennas and Propagation, 40, 8–12, 1982. Bouchard P., “Observations of attenuation due to liquid bearing stratocumulus clouds over Ottawa using a ground-based profiling radiometer,” 10th Specialist Meeting on Microwave Radiometry and Remote Sensing of the Environment, 11–14 March 2008, Florence, Italy. 16
Chakraborty, S., and M. Animesh (2012). A Comparative Study of Cloud Liquid Water Content from Radiosonde Data at a Tropical Location, Int. J. Geosci., 3, 44-49. Das S, Chakraborty S and Maitra A, Radiometric measurements of cloud attenuation at a tropical location in India, Journal of Atmospheric and Solar-Terrestrial Physics, 105-106, 2013, pp 97-100, 2013. Dissanayake, A., J. Allnutt, and F. Haidara, Cloud attenuation modeling for SHF and EHF applications, Int. J. Satell. Commun., 19, 335–345, 2001. Falayi E.O and Rabiu A.B, Estimation of global solar radiation using cloud cover and surface temperature in some selected cities in Nigeria. Archives of Physics Research, 2(3):99-109, 2011. Falayi E.O, The Impact of Cloud Cover, Relative Humidity, Temperature and Rainfall on Solar Radiation in Nigeria, Energy and Power, 3(6): 119-127, 2013. Garcia, P., A. Benarroch, and J. M. Riera, spatial distribution of cloud cover, Int. J. Satell. Commun. Network., 26, 141–155, 2008. Hall M.P.M, Effect of the troposphere on radio communication, IEEE Electromagnetic Wave series, Uk and USA, 1980. Ippolito L.J , Satellite communication systems engineering, atmospheric effects, satellite link design and system performance, Willey, UK , 2008 ITU-R p.840-5, Attenuation due to clouds and fog, ITU-R Recommendation, Geneva, Switzerland, 2012 ITU-R P. 676-10, Attenuation by atmospheric gases, ITU-R Recommendation Geneva, Switzerland, 2013. Liebe, H. J., An atmospheric millimeter wave propagation model, Int. J. Infrared Millimeter Waves, 10(6), 631–650, 1989. King M. D, Platnick S, Menzel W. P, Steven A. A, and Paul A. H, Spatial and Temporal Distribution of Clouds Observed by MODIS Onboard the Terra and Aqua Satellites, IEEE Trans On Geosc and Remote Sensing, 51(7), 3826-3852, 2013. Kumar A. and Sarkar S. K, Cloud attenuation and cloud noise temperature over some Indian eastern stations for Satellite communication, Indian Journal of Radio and Space Physics, 36, 375378, 2007. Mandeep, J. S., and S. I. S. Hassan, Cloud attenuation in millimeter wave and microwave frequencies for satellite applications over equatorial climate, Int. J. Infrared Millimeter Waves, 29, 201–206, 2008.
17
Ogunjobi K.O, Kim Y.J, Adedokun, J.A, Ryu S.Y and J.E Kim, Analysis of sky condition using solar radiation data at Kwangju and Seoul, South Korea and Ile-Ife, Nigeria, Theoretical and Applied Climatology, 72, 265-272, 2002. Omotosho T.V and E.B. Babatunde, Study of Cloud Impact on Fixed Satellite Link at Ku, Ka and V Bands in Nigeria, Australian Journal of Basic and Sciences, 4(8): 3287-3298, 2010.
Communication Applied
Omotosho T. V, Mandeep J S, and Abdullah M, Cloud-Cover Statistics and Cloud Attenuation at Ka- and V-Bands for Satellite Systems Design in Tropical Wet Climate, IEEE Antennas and Wireless Propagation Letters, 10, 1194-1196, 2011. Sarkar S. K, Ahmad I, Das J and A. K. De, Cloud height, cloud temperature and cloud attenuation in microwave and millimeter wave frequency bands over Indian Tropical East Cost. Int. J. Infrared Milli. Waves 26, 329–340, 2006 Sarkar S. K., Anil Kumar, Ahmed, I and M. M. Gupta (2006), Cloud morphology over Indian Tropical stations for Earth Space Communications, Int. J. Infrared Milli. Waves 27, 1005– 1017, 2006. Sarkar S. K, and A. Kumar (2007), Recent studies on cloud and precipitation phenomena for propagation characteristics over India, Indian Journal of Radio and Space Physics, 36, 502515, 2007 Slobin, S. D, Microwave noise temperature and attenuation of clouds: Statistics of these effects at various sites in the United States, Alaska and Hawaii, Radio Sci., 17, 1443–1454, 1982.
18
Highlights Geo-spatial distribution of cloud cover and noise temperature on satellite links. Mapping of correlation coefficients and dependence index of cloud cover Earth-space link budgeting for the proposed Nigerian multi-sensors satellite
19
S0273-1177(17)30179-5 http://dx.doi.org/10.1016/j.asr.2017.03.006 JASR 13140
To appear in:
Advances in Space Research
Received Date: Revised Date: Accepted Date:
13 September 2016 27 February 2017 8 March 2017
Please cite this article as: Ojo, J.S., Geo-spatial distribution of cloud cover and influence of cloud induced attenuation and noise temperature on satellite signal propagation over Nigeria, Advances in Space Research (2017), doi: http:// dx.doi.org/10.1016/j.asr.2017.03.006
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
GEO-SPATIAL DISTRIBUTION OF CLOUD COVER AND INFLUENCE OF CLOUD INDUCED ATTENUATION AND NOISE TEMPERATURE ON SATELLITE SIGNAL PROPAGATION OVER NIGERIA Joseph Sunday OJO Department of Physics, Federal University of Technology, Akure, P.M. B 704, Akure, Nigeria. Email: [email protected] or [email protected]
Abstract The study of the influence of cloud cover on satellite propagation links is becoming more demanding due to the requirement of larger bandwidth for different satellite applications. Cloud attenuation is one of the major factors to consider for optimum performance of Ka/V and other higher frequency bands. In this paper, the geo-spatial distribution of cloud coverage over some chosen stations in Nigeria has been considered. The substantial scale spatial dispersion of cloud cover based on synoptic meteorological data and the possible impact on satellite communication links at higher frequency bands was also investigated. The investigation was based on 5 years (2008 – 2012) achieved cloud cover data collected by the Nigerian Meteorological Agency (NIMET) Federal Ministry of Aviation, Oshodi Lagos over four synoptic hours of the day covering day and night. The performances of satellite signals as they traverse through the cloud and cloud noise temperature at different seasons and over different hours of days at Ku/W-bands frequency are also examined. The overall result shows that the additional total atmospheric noise temperature due to the clear air effect and the noise temperature from the cloud reduces the signal-to-noise ratio of the satellite receiver systems, leading to more signal loss and if not adequately taken care of may lead to significant outage. The present results will be useful for Earth-space link budgeting, especially for the proposed multi-sensors communication satellite systems in Nigeria. Keywords: Geo-spatial; Cloud cover; Cloud Attenuation; Cloud noise temperature; Satellite applications; Ku/W-bands; Nigeria 1. Introduction
The technological world is indeed yearning for more transformation on a daily basis and therefore calling for more bandwidth for services that employ radio and satellite telecommunication systems. Most of the new technological devices operating at higher frequency bands enjoy the benefit of higher data transmission rate, smaller antenna size, higher throughput and larger band spectrum to mention but few (Hall, 1980. Ippolito, 2008). However, since the migration from C to higher frequency bands, satellite and terrestrial signal impairment have been the major issues to be solved by the system designers. Studies revealed that satellite communication systems operating at frequency > 10 GHz are more susceptible to rain fade while cloud attenuation has been identified to be one of the major factors that can reduce the usage of Ku/V or higher frequency bands (Allnut and Rogers, 1989, Dissanayake et al., 2001, Sarkah et al., 2005, Das et al., 2013, Ali et al., 2014). Although, rain-induced attenuation plays a major role as far as signal degradation is concerned at these frequency ranges, the effect of cloud attenuation is relatively smaller at Ku and Ka frequency bands. However, at these frequency bands, Earth to satellite links are bound to experience some level of losses due to cloud at any period of the day; depending on the season and the climate of the region (Bouchard, 2008, Ali et al., 2013). Cloud attenuation is even more pronounced at other higher frequency bands, especially those operating at low availability satellite links. This is as a 1
result of the small terminals associated with these bands and the higher probability of occurrence of cloud when compared to rain attenuation. Cloud is known to comprise of tiny water droplets condensing out of the air to form liquid water. In the tropical region, Nigeria inclusive, the occurrence of cloud cover is more prevalent in the year due to the frequency of rainfall in this region when compared to the temperate region. For example, Nigeria climate comprises of about 70% as a yearly average of occurrence of rainy months (some part may sometime experience continuous rainfall for weeks) and thereby liable to more period of cloud cover and associated cloud attenuation (Slobin, 1982, Omotosho and Babatunde. 2010, Ali et al., 2014). Hence, it is paramount to examine the influence of cloud cover and the temperature within the cloud for accurate measurement of cloud-induced attenuation in Nigeria, especially for the design of low fade margin Earth-satellite communication links operating at Ku/V-bands and other higher frequencies. In spite of the attention that the effect of cloud cover has drawn, especially on Earth-satellite communication links at higher frequencies, the subject is just receiving attention in Nigeria. Other applications of cloud cover have always been directed to solar radiation (Ogunjobi et al., 2002, Augustine and Nnabuchi, 2009, Falayi and Rabiu, 2011, Falayi, 2013). For radio and satellite application, for example, Omotosho and Babatunde (2010) examined the impact of cloud on the fixed satellite communication link over some selected stations in Nigeria using satellite data. However, the stations were only categorized on regional basis and not on climatic basis. Also, Omotosho et al. (2011) reported on the effect of cloud cover and cloud attenuation at Ka/V-band frequencies for satellite systems application in tropical wet climate, however the result obtained was too generalized and cannot be specifically applied to geo-climatic region in Nigeria. The seasonal influence, a very strong factor to be considered, especially when determining the level of signal degradation that can be encountered during the worst months was not considered and the effect of noise within the cloud was also not examined. In addition, the attenuation due to cloud obtained was limited to 50 GHz frequency. The significance of the present paper is that it is the first statistical approach that will address the aforementioned shortcomings among others, using a more recent low cloud data obtained from the Nigerian meteorological stations. 2. Data details and Analysis
Daily cloud cover data obtained by the Nigerian Meteorological Agency (NIMET) Federal Ministry of Aviation, Oshodi Lagos, Nigeria from January 2008 to December 2012, have been used in the present study. Micropulse Lidar like conventional radar was used in each of the sites. The instrument comprises of a modified 8-inch telescope with bursts from a laser into the sky directly above the instrument. As the pulses emanating from the laser light travel through the sky, they may come across some tiny water droplets which include a cloud or aerosol particles. The scattered pulses reflect away from the particles and some are redirected back to the instrument below. Based on the time it takes the scattered and reflected pulses to return to the surface, the height of a cloud base can be accurately determined. Usually, the data are collected twice a day, during the day (0.00 hrs and 06.00 hrs) and during the night (12.00 hrs and 18.00 hrs) local time (LT). The low cloud amount is measured in eighths having a range from 0 (no cloud) to 8 (complete cover of sky). The synoptic data values recorded at NIMET stations were available for about 95% of the time. The missing or errors in the data were less than 5%. The cloud data obtained from NIMET have advantage of the availability of a larger number of stations in Nigeria over a large period rather than extrapolating data for a specific station using satellite data. Ten meteorological stations covering some selected cities in Nigeria were used with each of the sites located widely across Nigeria. The sites are considered as representative of different climatic regions (Rain forest belt-RFB, Sudan savannah-SS, Wooden belt-WB, Guinea Sudan-GS and Sahel-S) in Nigeria. Table 1 presents the 2
characteristics of the study sites. Other atmospheric parameters measured alongside the cloud cover are pressure, temperature and dew point temperature at different heights. In order to examine signal attenuation due to cloud, knowledge about the statistics of these clouds for an extensive period of time is required. Cloud is known to exist in different forms such as the, Cumulus (Cu), Stratocumulus (Sc), Cumulonimbus (Cb), Nimbostratus (Ns), Stratus (St), Altostratus (As) and so on. Each of this cloud type has their peculiar characteristics in terms of shape, effective radius, moisture content and liquid water contents among others. Detailed description is not discussed in this report due to paucity of space, but they are readily available in the literature (Slobin, 1982, Sarkah and Kumar, 2007, Garcia et al., 2008, King et al., 2013). Table 1: Characteristics of the study sites Climatic Region
Stations
Coordinates (Long oN, Lat oE)
Altitude (m)
Rain Forest Belt (RFB)
Calabar
4.9757, 8.3417
32
Uyo
5.0377, 7.9128
50
Benin
6.3350, 5.6037
88
Asaba
6.2059, 6.6959
52
Ibadan
7.3775, 3.9470
230
Makurdi
7.7322, 8.5391
114
Yola
9.2035, 12.4954
599
Minna
9.5836, 6.5463
350
Damaturu
11.7470, 11.9662
456
Sokoto
13.0059, 5.2476
356
Wooded Belt (WB)
Guinea Sudan (GS)
Sudan-Sahel (SS)
Sahel (S)
For the spatial distribution of cloud cover, the procedure stated in the work of Garcia et al. (2008) was adopted. Cloud cover is usually expressed in terms of octas, a unit that represents one-eighth of the sky vault. The sum of solid angles Ω, on the sky dome as depicted by the cloud cover varies from 0 (in case of clear sky) to Ωo (in case of overcast sky). Cloud cover amount (or cloud fraction) is then the ratio Ω/ Ωo, which may also vary between 0 and 1, measured in octas (Garcia et al., 2008). The samples of the time series derived from the synoptic meteorological data have been estimated based on two different approaches that correspond to the following modes: Mode 1: Each sample, Xn, of the time series x is defined as the amount of the sky vault covered by a specific type of cloud at the period of measurements, represented as: Xn =
On 8
(1)
where On = 0,1,2,….,8 is the number of octas in the nth data record. The fraction of the cloud cover obtained using equation (1) can be applied in meteorology. Mode 2: For a specific cloud type, the binary time series x is defined as:
X
n
=
O n ≥Ot 1 if 0 if O < O n t
{
(2)
3
where On is the number of octas in the nth data record (or time interval) and Ot is a threshold number of octas, Ot = 1,2,….8. Equation (2) provides information about the threshold level of the number of octas. For the spatial distribution of cloud cover, we have adopted the occurrence of cloud in pairs of sites at each of the climatic regions in Nigeria. This could be achieved by using the statistical metric measure like correlated coefficient, statistical dependence index and the joint probability of simultaneous occurrence of cloud of a specific type over each of the regions (Garcia et al., 2008). As defined in the work of Barbaliscia et al. (1992) and Garcia et al. (2008), if there exists two sites Q and R in the same region, the correlated coefficient can be estimated as: cov ( x , y ) ρ= (3) σ x ,σ y where x and y are the time series earlier defined in equations (1) and (2) for sites Q and R, respectively. The statistical dependence indexχQR, is also expressed as:
χ QR =
PQR
(4)
(P .P ) Q
R
where PQ and PR are the individual probabilities of cloud cover of a specific type at the sites Q and R, respectively. The joint probability of the simultaneous occurrence of cloud covers of a specific type, in the nth record, PQR is given as:
PQR =
1 n ∑ ( X n .Yn ) N n =1
(5)
where Xn and Yn are same as defined in equations (1) and (2). The systematic formulation of attenuation due to cloud at microwave and millimeter wave frequency follows nearly the same method as the method used in predicting attenuation due to rain. The significant difference between the two atmospheric parameters (rain and clouds) is that clouds are made up of suspended mist of water drops with very small diameters less than 1µm while rain is the liquid water in the form of droplets that have condensed from atmospheric water vapor and become heavy enough to fall under gravity (Hall, 1980, Ali et al., 2014). The degradation or impairment due to the suspended water droplets contained in atmospheric clouds can be estimated with some level of accuracy, using the model developed by Liebe (1989) for fog and cloud attenuation at frequencies up to 1000 GHz based on the Rayleigh scattering of electromagnetic wave, which is associated with a double-Debye model for the dielectric permittivity of water. The model has generally been adopted by ITU-R Recommendation ITU-R P.840-5 (2012). This same procedure is adopted in this present study as expressed step by step below: The expression for the specific attenuation within a cloud or fog, γc can be written as:
γ c = Kl M
(6)
M which is the liquid water density in the cloud (gm-3) is obtained from the radiosonde profiles at different NIMET stations. For the purpose of this work, M is estimated using the Salonen’s model from radiosonde data at each of the selected stations. Cloud’s formation is based on the knowledge of occurrence of high relative humidity whereby the radiosonde data show the evidence of LWC based on whether the relative humidity has gone beyond its critical value. According to Karsten’s 4
model, the formation of cloud occurs when the relative humidity exceeds 95% and the phase is determined by the temperature profile (Chakraborty and Animesh, 2012). Therefore, by using the thermodynamic concept, the LWC for each height level can be estimated from the following expression: Cp (Γd − Γs )dz M ad ( h ) = ∫ ρ (z ) (7) Lh where: Mad(h)= liquid water content under adiabatic condition; Lh = latent heat of vaporization given as 80 cal/gm; Cp = specific heat at constant pressure with the value 1.0035 J·g–1·k–1; ρ(z) = air density; Γd = dry adiabatic lapse rate; Γs = moist adiabatic lapse and varies from 4˚C/km to 9.8˚C/km depending on the seasonal variation of temperature. Hence, the adiabatic condition produces a maximum value of the LWC that is given by equation (8) as a result of the circulation of air mass accompanied by precipitation and freezing temperatures.
M = M ad (1.239 − 0.145 ln ∆h ) (8) where ∆h = height above the cloud base. M is then calculated at each pressure level at a particular radiosonde ascent. Integrating the LWC profile over height, the total value of the LWC is obtained at each ascent (Chakraborty and Animesh 2012). Also, the parameter Kl in equation (6) is the specific attenuation coefficient (dBkm-1)/gm-3) and based on Rayleigh scattering can be expressed as:
Kl =
0.819 f (dBkm-1) 2 ε ′′(1 + η )
(9)
where f is the frequency (GHz) , and
η=
2+ε′ ε ′′
(10)
where ε ′ and ε ′′ are the complex dielectric permittivity of water as a function of frequency and temperature. The parameters can be obtained using double-Debye model (Das et al., 2013). To obtain the attenuation due to clouds for a given probability of time, the statistics of the total columnar content of liquid water L (kg/m2) or, equivalently, mm of precipitable water for a given site must be known yielding: A=
LK l sin θ
(dB) for 90o > θ > 5o
(11)
where θ is the elevation angle and Kl is given as in equations (7) - (15) and The total atmospheric noise temperature Ttatm due to clear air effect (addition of Oxygen and water vapour) and the noise temperature from the cloud for different cloud thickness up to 2 km are also estimated for the selected stations during the worst month’s condition using (Slobin, 1982; Sarkar and Kumar, 2007): 1 Ttatm = T f 1 − (12) Ls and Ls = 10 [αtotal ( dB ) ] / 10 (13)
5
where Tf is the physical temperature, Ls is the loss factor, and αtotal is the total attenuation of radio wave due to oxygen, water vapour and cloud. For the purpose of this work, the attenuations of radio wave due to oxygen and water vapour are estimated using ITU-R P. 676-10 (2013) and processed with Matlab 2016a gaspl function for the selected frequencies over each of the study locations. 3. Results and discussion This section discusses in detail results on the morphology of cloud occurrences, spatial distribution of cloud cover, noise temperature within the cloud and their possible effect on Earth-satellite signal links at higher frequencies over the selected sites in Nigeria.
3.1. Cloud cover morphology for the selected sites Cloud occurrence morphology is necessary to understand the extent of degradation on signal traversing through Earth-satellite path and for optimizing power sharing in multi-sensors satellite systems. Clouds are known to consist of liquid water particles having diameters from about 0.1 to about 1.0 gm-3, while a typical radius of water droplets within non-rainy clouds varies between about 1 µm and 30 µm with a maximum droplet density of about 3-6 µm (Ippolito, 2008, Mandeep and Hassan, 2008). A particular cloud type will have a range of water particle sizes, while the thickness of the cloud varies from about 1.5 to 2.5 km depending upon the type of the cloud as reported by Kumar and Sarkar, (2007). As earlier reported in section 2, there exist several types of cloud, among which Cu/Cb are very important in Nigeria due to their characteristics. Cumulus clouds are low-level cloud with flat bottoms and rounded tops, and grow vertically. It is known to contain particles with diameters from about 4 to 15 µm; cumulonimbus clouds have particle diameters from 2 up to 100 µm, where the distinction between cloud particles and suspended rain is not clear. Cumulus cloud types with large vertical growth generally have large liquid water content while cumulonimbus results from when enough atmospheric instability, moisture, and lift are present. That often leads to a strong updraft within the cumulus cloud resulting to a mature and deep cloud. Cloud electrification occurs within the cumulonimbus clouds due to many collisions between charged water droplets, graupel (ice-water mix), and ice crystal particles, resulting in lightning and thunderstorm which is one of the characteristics of tropical Nigerian climate. Based on the local data obtained from NIMET, Table 2 presents the cloud cover for individual sites. The time-series have been calculated for each of the modes earlier described in equations (1) and (2). The time-averaged cloud amount is calculated by summing the samples for each type of cloud and dividing by the total number of synoptic records. The results from Table 2 revealed the climatic peculiarities expected at each site. It is seen that the occurrence of Cu/Cb cloud types is the more frequent in the synoptic records, the corresponding values in the table are relatively close to the values of the “low cloud” group. The larger cloud amounts are observed in southern regions within the rain forest climatic zone (Calabar, 52.6%), whereas sites located in the northern region of Nigeria gave the smaller cloud cover amounts (Damaturu, 14.9%). Ns presence is also more common in the Southern regions often affected by rainy fronts, which agree with the fact that stratiform rains are usually associated with Ns clouds. The present result do not agree with those earlier obtained in Spain (Temperate region) where a higher percentage of cloud cover occurs in the northern regions with St/Sc cloud type being the most frequent in the synoptic records (Gracia et al., 2008). This shows the dynamical nature of cloud cover and how geographical location can dictate the occurrence of cloud cover. A more random distribution is noted in the case of St or Sc clouds. The values in the right-hand side of Table 2, give an indication of how often the threshold “no cloud” has been exceeded. The Southern region has a close proximity to the Sea, (in this case the Atlantic Ocean) which produces more cloud as earlier reported on the work of Garcia et al. (2008). There is over 70% occurrence of cloud 6
cover in the Southern regions. Also, the most global variance in cloud radiative response between Global Cloud Models (GCMs) is due to low clouds with about 50% due to the cumulus regime and 20% due to the regime characterized by clouds undergoing transition from cumulus to cumulonimbus. Table 2. Cloud cover for individual sites Mode 1
Mode 2
Climatic Region
Stations
Low (%)
Cu/Cb (%)
St/Sc (%)
Ns (%)
Low (%)
Cu/Cb (%)
St/Sc (%)
Ns (%)
RFB
Calabar
52.1
44.1
5.0
7.9
73.0
49.5
14.8
9.4
Uyo
51.6
27.8
3.2
7.7
70.2
56.2
13.5
5.8
Benin
41..8
21.8
2.9
5.8
41.0
30.0
6.6
3.3
Asaba
42.2
33.3
4.0
6.3
50.8
36.8
7.2
4.9
Ibadan
41.8
31.9
3.9
6.0
50.1
33.4
7.1
3.6
Makurdi
38.5
7.0
1.3
2.1
46.6
39.9
5.5
2.8
Yola
33.6
26.6
3.2
5.0
42.5
37.7
8.7
8.2
Minna
32.1
19.8
2.2
4.0
38.6
30.4
6.5
4.9
Damaturu
14.9
10.2
5.6
5.2
19.3
10.7
8.5
1.1
Sokoto
18.9
12.0
1.5
3.6
25.8
19.3
3.4
1.6
B
GS
SS
S
The cloud cover observations over the four times of the day comprising the daytime and the nighttime (06:00, 12:00, 18:00 and 00:00 hrs LT) are also characterized and results are presented in Figs. 1 (a-b) for RFB sites, Figs 2 (a-b) for WB sites, Figs 3 (a-b) for GS sites, Figs 4 (a-b) for SS sites and Figs. 5 (a-b) for S sites during daytime and the nighttime respectively. For ease of result analysis, the sites are paired alongside the climatic region as earlier categorized in section 2. For example Figs. 1 (a-b) present the monthly distribution of cloud cover over RFB region during daytime and nighttime respectively. In Fig. 1(a) for example, the result shows that over the RFB region, daytime low cloud is significant in the months of June, July, August and September at 6 hrs for 19, 21, 24 and 22 days and at 12 hrs LT for 18, 19, 21 and 19 days in Calabar while daytime low cloud is found to occur in 15, 17, 22 and 18 days and 14, 14, 16 and 18 days in Uyo during the respective hours and same months. It is however noticed that, irrespective of the months, daytime low cloud covers occurred more frequently in the 6 hrs LT when compared with the 12 hrs LT. The result for the nighttime low cloud as presented in Fig. 1 (b) also shows that, over the RFB coastline, the sky is partially covered with clouds in the months of April, June, July, August, September and October for 16, 19, 19, 21, 15 and 16 days in Calabar and 14, 16, 15, 19, 16 and 14 days in Uyo during midnight (12 hrs LT). Further results show that at the 18 hrs LT, the sky is covered with clouds for 11, 9, 10 and 11 days in the months of April, August, September and 7
October respectively and 9, 5 and 6 days in the months of April, August and October respectively in Calabar. It is evident here that the effect of ocean breeze is more influenced in Uyo in the nighttime when compared with Calabar. It could also be seen that Figs 1(b) and 3 (b) have a variation compared to some extent to Figs 1(a) and 3(a), this is may be due some similar climatic features between the two climatic regions (Rain forest belt and Guinea Sudan). For example, both regions have a moderate occurrence of the average rainfall amount due to cloud-cover when compared with other climatic regions considered in this work. Other climatic regions follow the same trend, although with different numbers of days the sky is covered with clouds and probably different months of the year as presented in Figs. 2 – 5. Uyo (6 hrs LT) Average (6 hrs LT) Calabar (12 hrs LT)
25 NUMBER OF DAYS
NUMBER OF DAYS
25
Calabar (6 hrs LT) Uyo (12 hrs LT) Average (12 hrs LT)
20 15 10 5
Uyo (18 hrs LT) Average (18 hrs LT) Calabar (00 hrs LT)
Calabar (18 hrs LT) Uyo (00 hrs LT) Average (00 hrs LT)
20 15 10 5
0
0 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec MONTH
MONTH
Fig. 1: Monthly distribution of cloud coverage over RFB region during (a) daytime and (b) nighttime Benin (6 hrs LT) Average (6 hrs LT) Asaba (12 hrs LT)
25 NUMBER OF DAYS
NUMBER OF DAYS
25
Asaba (6 hrs LT) Benin (12 hrs LT) Average (12 hrs LT)
20 15 10 5
Benin (18 hrs LT) Average (18 hrs LT) Asaba (00 hrs LT)
Asaba (18 hrs LT) Benin (00 hrs LT) Average (00 hrs LT)
20 15 10 5 0
0 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec MONTH
MONTH
Fig. 2: Monthly distribution of cloud coverage over WB region during (a) daytime and (b) nighttime Ibadan (6 hrs LT) Average (6 hrs LT) Makurdi (12 hrs LT)
20
25
Ibadan (18 hrs LT) Average (18 hrs LT) Makurdi (00 hrs LT)
Makurdi (18 hrs LT) Ibadan (00 hrs LT) Average (00 hrs LT)
NUMBER OF DAYS
20
15
15
10
10
5
5 0
Dec
Nov
Oct
Sep
Aug
Jul
Jun
May
Apr
Feb
Mar
0 Jan
NUMBER OF DAYS
25
Makurdi (6 hrs LT) Ibadan (12 hrs LT) Average (12 hrs LT)
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec MONTH
MONTH
Fig. 3: Monthly distribution of cloud coverage over GS region during (a) daytime and (b) nighttime. 8
Yola (6 hrs LT) Minna (12 hrs LT) Average (12 hrs LT)
NUMBER OF DAYS
NUMBER OF DAYS
18 16 14 12 10 8 6 4 2 0
Minna (6 hrs LT) Average (6 hrs LT) Yola (12 hrs LT)
18 16 14 12 10 8 6 4 2 0
Minna (18 hrs LT) Average (18 hrs LT) Yola (00 hrs LT)
Yola (18 hrs LT) Minna (00 hrs LT) Average (00 hrs LT)
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
MONTH
MONTH
18 16 14 12 10 8 6 4 2 0
Damaturu (6 hrs LT) Average (6 hrs LT) Sokoto (12 hrs LT)
Sokoto (6 hrs LT) Damaturu (12 hrs LT) Average (12 hrs LT)
NUMBER OF DAYS
NUMBER OF DAYS
Fig. 4: Monthly distribution of cloud coverage over SS region during (a) daytime and (b) nighttime
18 16 14 12 10 8 6 4 2 0
Damaturu (18 hrs LT) Average (18 hrs LT) Sokoto (00 hrs LT)
Sokoto (18 hrs LT) Damaturu (00 hrs LT) Average (00 hrs LT)
MONTH
MONTH
Fig. 5: Monthly distribution of cloud coverage over S region dur during ing (a) daytime and (b) nighttime
3.2.Statistical evaluation of the spatial distribution of cloud cover To understand the spatial distribution and occurrence of cloud cover, we have adopted the sites pairing method using the correlation coefficient and statistical dependence index based on the climatic region in Nigeria. Results are presented in the form of maps to give distinct information needed for link behavior of cloud cover on each of the sites. The maps are developed using Kriging Interpolation Technique (KIT) by incorporating all the statistical results using equations (3) –(6). KIT was used for the spatial interpolation of correlation coefficient and statistical dependence index cloud cover values based on regular grid. This will assist to generate a highly consistent and predictable inter-site cloud cover variation for any of the geographical location within the region under investigation The contour maps are developed based on the results presented in each of the columns in Table 2. Fig. 6 (a) presents the map for mode 1, for the low cloud group of the cloud cover amount over the regions. If mode 2 is used, a threshold of 1 octa should be considered and the probability of detecting any amount of cloud in any sky region for the low cloud group can also be obtained. The result is plotted on the map represented in Fig. 6(b). Each of the maps has different shape of the contour lines based on the climatic characteristics of each of the geographical region. For example the rainforest zone, which is close to the sea produces relatively large difference in cloud cover between sites with closer proximity. This is evident in the 9
cloud cover values of about 50% in Calabar near the sea shore which reduces below 40% for Asaba site moving upward the sea thereby creating a steep gradient in contour lines. The map derived using mode 2 with a threshold of 1 octa shows an increase in cloud detection in coastal and mountainous areas. This may be often due to cumulus cloud reports, which may be visible in small sky regions in the RFB and WB sites. The cloud cover amount reduces as we moved towards the northern part of the country. (%)
12
(%)
12
80
11
11
52
75 48
10
10
70
9
40 36
8
L A T IT U D E (D E G )
L A T IT U D E (D E G )
44
65
9
60 55
8
50
32 7
7
45
28 6
24
40
6
35 20
30
5
5
16 4 5
6
7
8
9
10
11
12
13
12
LONGITUDE (DEG)
25 4 5
6
7
8
9
10
11
12
13
20
LONGITUDE (DEG)
(a)
(b)
Fig. 6: Map of cloud cover value for the low cloud group using (a) mode 1 and (b) mode 2. Figures 7 (a) and (b) also show the map of the correlation coefficients and the dependence index respectively for the various stations. The contour map is plotted for the sites by taking into account the values relative to the inter-site under investigation. The value for these parameters diminishes with distance as we move from the southern coastline to the northern plain sites; the shape of the contour lines is strongly influenced by climatic and geographical factors of the selected sites. This is evident in the contour lines, because if only the influence of distance is considered, the contour will just be circling lines. For example the correlation coefficient maps depicted in Fig. 7(a) shows that the contour lines expand along the RFB coastline (Calabar and Uyo) but compressed towards the GS which is more mountainous topographically and enlarging towards the relatively flat northern part of the country (SS and S). The contour lines around the WB region (Asaba and Benin) tend to follow the RFB bearing the narrow coastal strip. These trends are also observed in the statistical dependence index (χ) map as presented in Fig. 7 (b). The implication of the maps is in the application of the geo-spatial distribution analysis in the design of the multi-sensors satellite systems over the study sites and application of the site diversity techniques among others. For example the information from the maps will enable communication system designers to cater for the level of signal degradation in terms of sharing the available power (downlink or uplink) to mitigate the effects of simultaneous bad weather in the given section of the coverage area. As noted in the work of Garcia et al. (2008), if a given site X1 needed additional power due to cloud attenuation, the statistical dependence index values estimated for the pair of sites X1 –X2 will show the probability that additional power is also needed in the link to site X2. Also, information from the maps will assist to know the level of site dependency as well as the degree of relation between sites. For example, if χ is equal to or approaches 1, statistical independence is achieved and both sites can be dealt with independently. However, higher values of
10
χ indicate that both sites are prone to cloud attenuation. The same trends are applicable to the correlation coefficients (ρ) bearing in mind the diverse range of ρ. 12
12
11
11
2
0.95
10
10
1.9
0.9
1.8
0.8 0.75
8
0.7 0.65
7
0.6
LATITUDE (DEG)
LATITUDE (DEG)
0.85 9
9
1.7 1.6 1.5
8
1.4 1.3
7
1.2
0.55 0.5
6
1.1 6
1
0.45 0.4
5
0.9 5
0.8
0.35 0.3
4 5
6
7
8
9
10
11
12
0.7 0.6
4 5
13
6
7
8
9
10
11
12
13
LONGITUDE (DEG)
LONGITUDE (DEG
(a)
(b)
Fig. 7: Statistical map of (a) correlation coefficients and (b) dependence index relatives to the study sites 3.3.Specific attenuation and total attenuation due to cloud coverage
The specific attenuation due to cloud has been estimated for each of the sites at different climatic region based on ITU-R P.840-5 (2012) method, using equations (6) – (10) on the basis of the liquid water content 1 gm-3 and for a cloud temperature of 273 K as input parameters. The estimations are done over frequency ranges from the Ku to W-band following the IEEE Standard 521-2002, namely: 12, 20, 30, 50, 70 and 100 GHz. Results are presented in Table 3 for the specific attenuation on radio wave due to cloud cover in the selected sites across Nigeria. It is seen from Table 3 that the specific attenuation over the RFB coastline for water content ~ 1 gm-3 range between 0.02 and 5.32 dB/km at Ku – W band frequencies (12-100 GHz) for Calabar and between 0.18 – 4.89 dB/km at Uyo. It has earlier been noticed that the variation in the specific attenuation is due to the variation of cloud particle temperature (Sarkar et al., 2006). It is worth mentioning that the specific attenuation due to cloud decreases from the southern coastline region, which is characterized by more occurrence of cloud cover to the plain northern part of the country that is characterized by lesser cloud coverage. This is clearly evident in the result presented in Table 3 for all the stations at frequencies 12 – 100 GHz. Following the procedure highlighted in ITU-R P.840-5 (2012) using equations (6) – (11), the attenuation due to the cloud over the selected sites were estimated. The estimation was based on the total columnar content of liquid water L (kgm-2) as stated in equation (11). The NIGCOMSAT-R1 elevation angle of 42.5o has been used since it is an indigenous satellite for Nigerian communication visible over all the stations. The results of the total attenuation due to cloud at the specified frequency range for different cloud thickness of 1 km, 1.5 km and 2 km are presented in Table 4. The values have been chosen to conform to the low cloud thickness (Cu/Cb) associated with Nigeria climate. Generally speaking, the cloud attenuation observed over the locations show that the values are lower in Ku-band frequency and higher at W-band frequency. It is also evident from the results that cloud-induced attenuation increases as cloud thickness increases. The implication is that uplink frequencies will be grossly attenuated from the Earth-satellite paths even during clear sky conditions. The results, based on the climatic categorization and cloud thickness show that even at the same climatic region, mostly at V and W-band frequencies, the differences in 11
cloud attenuation value can be as high as 3 dB most especially at the RFB coastline even during clear air conditions. Some results earlier obtained by Omotosho and Babatunde (2010) were also corroborated in this study. For example, it was reported in the paper that cloud attenuation is generally higher in the coastal region than in the plain Northern region of Nigeria and that at Vband, the degradation may become a major concern for satellite links especially in the coastline region. The information presented in Table 4 will also assist in estimating atmospheric noise generated by the cloud which serves as additional signal degradation to Earth-satellite links even under clear weather conditions. 1.1. Enhanced noise temperature due to cloud
The total atmospheric noise temperature was also estimated to determine the degree of extra noise due to the noise generated by cloud and the consequences on satellite signal receivers in the region. The results on the atmospheric noise temperature with thickness 1, 1.5 and 2 km using equations (12) and (13) are presented in Figs. 8 (a-e) for RFB, WB, GS, SS and S region respectively. As usual, the sites are paired based on the climatic categorization. For example, in Fig 8 (a) for the RFB region, the total atmospheric noise temperature over Calabar from 1 to 2 km thickness varies between 15.37 and 24.57 K at 12 GHz Ku-band frequency, 63.22 and 87.63 K at 20 GHz K-band frequency, 85.58 and 110.45 K at 30 GHz Ka-band frequency, 199.43 and 286.64 K at 50 GHz Vband frequency, 406.67 and 459.56 K at 70 GHz and 648.34 and 741.52 K at W-band frequency. Uyo which also belongs to RFB region shares similar values of total atmospheric noise temperature from 1 to 2 km thickness and the values vary between 14.35 and 22.30 K at 12 GHz Ku-band frequency, 61.09 and 83.45 K at 20 GHz K-band frequency, 82.34 and 102.35 K at 30 GHz Kaband frequency, 198.34 and 272.36 K at 50 GHz V-band frequency, 404.26 and 457.63 K at 70 GHz and 643.41 and 736.23 K at W-band frequency. The trend continues in other climatic regions, although with different values of total atmospheric noise temperature. The implication of the results is that the additional total atmospheric noise temperature due to the clear air effect and the noise temperature from the cloud reduces the signal to noise ratio (SNR) of the satellite receiver systems, leading to more signal loss and if not adequately taken care of may lead to significant outage. The effect may even be more pronounced for low noise receiving satellite systems.
12
Table 3: Specific attenuation of radio wave due to cloud cover in Nigeria. Station
Calabar
Uyo
Asaba
Benin
Makurdi
Ibadan
Yola
Minna
Damaturu
Sokoto
Frequency (GHz) 12 20 30 50 70 100 12 20 30 50 70 100 12 20 30 50 70 100 12 20 30 50 70 100 12 20 30 50 70 100 12 20 30 50 70 100 12 20 30 50 70 100 12 20 30 50 70 100 12 20 30 50 70 100 12 20 30 50 70 100
Specific attenuation (dB/km) 0.22 0.42 1.34 2.65 4.24 5.35 0.18 0.38 1.11 2.31 4.01 4.89 0.15 0.36 0.89 1.89 3.91 4.04 0.11 0.32 0.81 1.53 3.75 4.02 0.16 0.37 0.83 1.63 3.67 4.03 0.14 0.28 0.82 1.59 3.23 3.89 0.08 0.23 0.67 1.49 3.18 3.83 0.06 0.21 0.58 1.41 3.07 3.76 0.05 0.2 0.49 1.38 3.02 3.29 0.06 0.18 0.47 1.22 3.01 3.19
13
Table 4: Cloud-induced attenuation of radio wave at different cloud thickness over Nigeria Station
Calabar
Uyo
Asaba
Benin
Makurdi
Ibadan
Yola
Minna
Damaturu
Sokoto
Frequency (GHz) 12 20 30 50 70 100 12 20 30 50 70 100 12 20 30 50 70 100 12 20 30 50 70 100 12 20 30 50 70 100 12 20 30 50 70 100 12 20 30 50 70 100 12 20 30 50 70 100 12 20 30 50 70 100 12 20 30 50 70 100
1.0 km 0.43 0.89 1.95 5.26 7.92 10.84 0.32 0.75 1.82 4.89 6.62 9.89 0.21 0.42 1.22 4.01 6.11 9.13 0.18 0.31 1.01 3.82 5.02 8.22 0.2 0.38 1.12 3.97 5.14 8.64 0.13 0.24 1.02 3.11 5.08 8.19 0.1 0.19 0.92 3.06 4.98 8.04 0.11 0.21 1.01 3.08 5.01 8.04 0.09 0.16 0.84 2.76 4.23 7.23 0.06 0.14 0.46 2.14 4.01 6.24
14
1.5 km 0.67 0.98 2.05 5.74 8.12 10.97 0.42 0.85 2.01 4.99 6.87 9.96 0.42 0.65 1.34 4.45 6.82 9.23 0.22 0.53 1.21 3.98 5.65 8.75 0.26 0.45 1.34 4.03 5.58 8.96 0.31 0.42 1.39 3.42 5.78 8.89 0.22 0.36 1.26 3.22 5.08 8.21 0.19 0.31 1.18 3.17 5.14 8.35 0.12 0.28 0.98 2.97 4.82 7.68 0.08 0.21 0.78 2.35 4.32 6.85
2.0 km 0.74 1.23 2.24 5.89 10.76 14.83 0.65 0.98 2.73 5.28 9.24 12.26 0.48 0.76 1.42 4.56 7.98 11.25 0.32 0.65 1.54 4.23 6.87 10.32 0.35 0.72 2.01 4.36 7.22 10.87 0.38 0.54 1.89 3.98 6.22 9.89 0.28 0.45 1.67 3.65 6.12 9.57 0.21 0.42 1.43 3.32 5.89 9.14 0.14 0.32 1.14 3.01 5.22 8.45 0.09 0.28 0.96 2.89 5.01 7.56
800
(a)
700 600 500 400 300 200 100 0
Asaba-1 km cloud thickness Asaba-1.5 km cloud thickess Asaba-2.0 km cloud thickness Benin- 1km cloud thickness Benin-1.5 km cloud thickess
600 500 400 300 200 100 0
0
50 100 FREQUENCY (GHz)
150
0
800
800
(c)
700
Mak-1 km cloud thickness Mak-1.5 km cloud thickess Mak-2.0 km cloud thickness Iba- 1km cloud thickness Iba-1.5 km cloud thickess
600 500 400 300 200 100 0
50 100 FREQUENCY (GHz)
Yola-1 km cloud thickness Yola-1.5 km cloud thickess Yola-2.0 km cloud thickness Minna- 1km cloud thickness Minna-1.5 km cloud thickess
600 500 400 300 200 100 0
50 100 FREQUENCY (GHz)
ATMOSPHERIC NOISE TEMPERATURE (K)
0
150
800 700 600 500 400 300 200 100 0
150
(d)
700 ATMOSPHERIC NOISE TEMPERATURE (K)
ATMOSPHERIC NOISE TEMPERATURE (K)
(b)
700
Cal-1 km cloud thickness Cal-1.5 km cloud thickess Cal-2.0 km cloud thickness Uyo- 1km cloud thickness Uyo-1.5 km cloud thickess
ATMOSPHERIC NOISE TEMPERATURE (K)
ATMOSPHERIC NOISE TEMPERATURE (K)
800
0
50 100 FREQUENCY (GHz)
150
(e) Dam-1 km cloud thickness Dam-1.5 km cloud thickess Dam-2.0 km cloud thickness Sok- 1km cloud thickness Sok-1.5 km cloud thickess
0
50 100 FREQUENCY (GHz)
150
Fig. 8: Total atmospheric noise temperature with thickness 1, 1.5 and 2 km for different frequencies over (a) RFB (b) WB (C) GS (d) SS and (e) S. Conclusion In this paper, 5 years (2008 – 2012) recent archived cloud cover data collected by the Nigerian Meteorological Agency over four synoptic hours of the day covering day and night have been used to evaluate the spatial distribution of cloud cover and the effect of cloud attenuation and cloud noise temperature on the performance of Earth-satellite links across the Geo-climatic region of Nigeria. The results on spatial distribution of cloud cover shows that the occurrence of Cu/Cb cloud type is the most frequent in the synoptic records, with the larger cloud amounts observed in southern regions with rain forest climatic zone while the sites located in the northern region of Nigeria gave 15
the smaller cloud cover amounts. The results on cloud cover observations over the four times of the day comprising the daytime and the nighttime also show that irrespective of the months, daytime low cloud covers occurred more frequently in the 6 hrs LT when compared with the 12 hrs LT while nighttime low cloud covers occurred more frequently in the 00 hrs LT when compared with the 18 hrs LT. Contour maps generated for the correlation coefficients and the relative dependence index will be useful in the design of the multi-sensor satellite systems over the study sites and application of the site diversity techniques among others. The results presented for the specific attenuation due to cloud revealed decreases in value from the southern coastline region, which is characterized by more occurrence of cloud cover to the plain northern part of the country that is characterized by lesser cloud coverage. Also, results on total attenuation shows that even in the same climatic region, mostly at V and W-bands frequency, the difference of cloud attenuation can be as high as 3 dB most especially at the RFB coastline even during clear air conditions. When the coastline is compared with plain northern sites, the difference can be as high as 7 dB at W-band frequency and 2 km cloud thickness. Results on additional total atmospheric noise temperature due to clear air effect and the noise temperature of the cloud lower the signal-to-noise-ratio of the satellite receiver systems, leading to more signal loss and if not adequately taken care of may lead to significant outage. Acknowledgements
The author wishes to acknowledge the Nigerian Meteorological Agency for the data used in this work and more especially for the International Centre for Theoretical Physics (ICTP), Trieste Italy for the latest software and access to some of the materials used. The effort of Professor O.S Ajayi to proof read the manuscript is also acknowledged. References
Ali Mohammed Al-Saegh, A. Sali, J. S. Mandeep, Alyani Ismail, Abdulmajeed H.J. Al-Jumaily and Chandima Gomes, Atmospheric Propagation Model for Satellite Communications, http://dx.doi.org/10.5772/58238, INTECH, 249-273, 2014. Ali Mohammed Al-Saegha, Aduwati Salia, J.S. Mandeep and Alyani Ismaila, Extracted atmospheric impairments on earth-sky signal quality in tropical regions at Ku-band, Journal of Atmospheric and Solar-Terrestrial Physics, 104, 96-105, 2013 Allnutt J. E, and D. V. Rogers, Low-fade-margin systems: propagation considerations and implementation approaches. IEE Conf. Publ. 301, 6–9, 1989. Augustine C and Nabuchi M.N, Correlation of cloudness index for four selected cities in Nigeria, The pacific Journal of science and Tech, 10(2), 568-573, 2009. Barbaliscia F, Ravaioli G, Paraboni A, Characteristics of the spatial statistical dependence of rainfall rates over large areas. IEEE Transactions on Antennas and Propagation, 40, 8–12, 1982. Bouchard P., “Observations of attenuation due to liquid bearing stratocumulus clouds over Ottawa using a ground-based profiling radiometer,” 10th Specialist Meeting on Microwave Radiometry and Remote Sensing of the Environment, 11–14 March 2008, Florence, Italy. 16
Chakraborty, S., and M. Animesh (2012). A Comparative Study of Cloud Liquid Water Content from Radiosonde Data at a Tropical Location, Int. J. Geosci., 3, 44-49. Das S, Chakraborty S and Maitra A, Radiometric measurements of cloud attenuation at a tropical location in India, Journal of Atmospheric and Solar-Terrestrial Physics, 105-106, 2013, pp 97-100, 2013. Dissanayake, A., J. Allnutt, and F. Haidara, Cloud attenuation modeling for SHF and EHF applications, Int. J. Satell. Commun., 19, 335–345, 2001. Falayi E.O and Rabiu A.B, Estimation of global solar radiation using cloud cover and surface temperature in some selected cities in Nigeria. Archives of Physics Research, 2(3):99-109, 2011. Falayi E.O, The Impact of Cloud Cover, Relative Humidity, Temperature and Rainfall on Solar Radiation in Nigeria, Energy and Power, 3(6): 119-127, 2013. Garcia, P., A. Benarroch, and J. M. Riera, spatial distribution of cloud cover, Int. J. Satell. Commun. Network., 26, 141–155, 2008. Hall M.P.M, Effect of the troposphere on radio communication, IEEE Electromagnetic Wave series, Uk and USA, 1980. Ippolito L.J , Satellite communication systems engineering, atmospheric effects, satellite link design and system performance, Willey, UK , 2008 ITU-R p.840-5, Attenuation due to clouds and fog, ITU-R Recommendation, Geneva, Switzerland, 2012 ITU-R P. 676-10, Attenuation by atmospheric gases, ITU-R Recommendation Geneva, Switzerland, 2013. Liebe, H. J., An atmospheric millimeter wave propagation model, Int. J. Infrared Millimeter Waves, 10(6), 631–650, 1989. King M. D, Platnick S, Menzel W. P, Steven A. A, and Paul A. H, Spatial and Temporal Distribution of Clouds Observed by MODIS Onboard the Terra and Aqua Satellites, IEEE Trans On Geosc and Remote Sensing, 51(7), 3826-3852, 2013. Kumar A. and Sarkar S. K, Cloud attenuation and cloud noise temperature over some Indian eastern stations for Satellite communication, Indian Journal of Radio and Space Physics, 36, 375378, 2007. Mandeep, J. S., and S. I. S. Hassan, Cloud attenuation in millimeter wave and microwave frequencies for satellite applications over equatorial climate, Int. J. Infrared Millimeter Waves, 29, 201–206, 2008.
17
Ogunjobi K.O, Kim Y.J, Adedokun, J.A, Ryu S.Y and J.E Kim, Analysis of sky condition using solar radiation data at Kwangju and Seoul, South Korea and Ile-Ife, Nigeria, Theoretical and Applied Climatology, 72, 265-272, 2002. Omotosho T.V and E.B. Babatunde, Study of Cloud Impact on Fixed Satellite Link at Ku, Ka and V Bands in Nigeria, Australian Journal of Basic and Sciences, 4(8): 3287-3298, 2010.
Communication Applied
Omotosho T. V, Mandeep J S, and Abdullah M, Cloud-Cover Statistics and Cloud Attenuation at Ka- and V-Bands for Satellite Systems Design in Tropical Wet Climate, IEEE Antennas and Wireless Propagation Letters, 10, 1194-1196, 2011. Sarkar S. K, Ahmad I, Das J and A. K. De, Cloud height, cloud temperature and cloud attenuation in microwave and millimeter wave frequency bands over Indian Tropical East Cost. Int. J. Infrared Milli. Waves 26, 329–340, 2006 Sarkar S. K., Anil Kumar, Ahmed, I and M. M. Gupta (2006), Cloud morphology over Indian Tropical stations for Earth Space Communications, Int. J. Infrared Milli. Waves 27, 1005– 1017, 2006. Sarkar S. K, and A. Kumar (2007), Recent studies on cloud and precipitation phenomena for propagation characteristics over India, Indian Journal of Radio and Space Physics, 36, 502515, 2007 Slobin, S. D, Microwave noise temperature and attenuation of clouds: Statistics of these effects at various sites in the United States, Alaska and Hawaii, Radio Sci., 17, 1443–1454, 1982.
18
Highlights Geo-spatial distribution of cloud cover and noise temperature on satellite links. Mapping of correlation coefficients and dependence index of cloud cover Earth-space link budgeting for the proposed Nigerian multi-sensors satellite
19