Climatology of anomalous propagation radar over Douala, Cameroon

Transcription

1 METEOROLOGICAL APPLICATIONS Meteorol. Appl. 21: (2014) Published online 18 April 2012 in Wiley Online Library (wileyonlinelibrary.com) DOI: /met.1321 A. Lenouo* Department of Physics, University of Douala, Cameroon ABSTRACT: The propagation of electromagnetic waves emitted from ground-based meteorological radar is determined using high-resolution radiosonde data. In this work, a detailed analysis of surface ducts has been undertaken to determine the anomalous propagation days on the coastal site of Douala in Cameroon (4 N, 9.7 E) over the Gulf of Guinea. The median duct strength shows that the strongest ducts are seen in the rainy season and that the surface duct occurs more frequently during the day than at night. The duct strength seasonal variability shows a value over Douala of about 7.2 M-units for 1200 UTC and 4.5 M-units for 0000 UTC in January, whereas in July the duct strengths are stronger during the day ( 12.8 M-units at 1200 UTC) than at night ( 9.8 M-units at 0000 UTC). The Cloudsat data products for Douala and neighbouring areas during the days of 26 January 2009 at UTC and 5 July 2009 at UTC were also analysed in association with the statistical discussion. KEY WORDS refractive index; anomalous propagation; rainfall; Gulf of Guinea Received 15 August 2011; Revised 12 January 2012; Accepted 6 March Introduction Weather radar is a radar pulse and is used in meteorology to identify the position, intensity and movement of precipitation. One of the main utilities of weather radar is to detect precipitation and approximate its intensity remotely (for example, for the inspection of the flow flood warning and other related purposes). Weather users need to know the amount of rain that falls over large areas. A radar network complements the rain gauge network because it covers a larger area: the former can be used to calibrate the latter. It becomes important to know the refractive index for atmospheric use of radar and other meteorological applications or telecommunications. In certain meteorological conditions, low-tilt beams emitted from groundbased radar can become trapped and can even be deflected toward the surface, leading to spurious backscattered signals and hence erroneous interpretation (e.g., a false precipitation echo). This phenomenon is referred to as anomalous propagation (AP), and the layer inside which the beam bends downward is called the duct. Ducting conditions have been widely investigated in various regions of the world using upper-air measurements. These include studying their variation with temperature (Dalmaz, 1977), on spatial and seasonal variability in Turkey (Celal Barla, 1986), on the conditions of the propagation of radio waves in Earth s atmosphere in Istanbul (4 N, 29 E) Turkey (Mentes and Kaymaz, 2007) or the temporal variability of the refractive index at 100 m from the ground, measured at Akure (7.15 N, 5.12 E), Nigeria (Falodum and Ajewole, 2006). For the last case specifically, the authors showed that seasonal variations in the vertical gradient of the index have extremes which correspond to the dry and rainy seasons in this city. The statistical study of duct conditions was also done at Wallops Island, Virginia, using radiosonde observations at high * Correspondence to: A. Lenouo, Faculty of Science, Department of Physics, University of Douala, Douala 00237, Cameroon. lenouo@yahoo.fr resolution, measured with a helicopter by Babin (1996), and in Barcelona, Spain (Bech et al., 2002). Climatological maps of the occurrence of the super-refracting and ducting layers, and maps of refractivity gradient, have been constructed (Lopez, 2009). A more recent study on refractive conditions by Mesnard and Sauvageot (2010) was undertaken in Bordeaux-Merignac, France, to determine the anomalous propagations at their radar site. The purpose of the present study is twofold. The first is to derive statistics of the surface ducts over Douala-Cameroon, which has important implications on the reliability of radar and communication systems. The second is to determine the characteristics of the surface ducts and refractivity profiles that result from seasonal variation of the climate in this part of the world, characterized by wet and dry seasons. The first step is analysis of the surface ducts occurring in both conditions and then an attempt to identify the meteorological differences using associated weather maps is undertaken. In the last section, results are discussed and conclusions presented. The results of this study are of great interest to both the civilian and military community in the Gulf of Guinea, as noted by Mentes and Kaymaz (2007). The coastal location and the proximity to the mountains makes the city of Douala very critical in this study. 2. Methodology and data 2.1. Methodology The refractive index of air is close to 1 and is used in meteorology co-index N defined by Bean and Dutton (1968), Steiner and Smith (2002) and Mentes and Kaymaz (2007): N = (n 1) 10 6 = 77.6 T ( P e T where e is water vapour (hpa), T the absolute temperature (K) and P the pressure (hpa). The effect of the Earth s curvature ) (1) 2012 Royal Meteorological Society

2 250 A. Lenouo (radius R) at an altitude z can be accounted for by considering a modified refractivity M (e.g., Lopez, 2009), defined as: height-finding radars, decreased/increased detection ranges and shortened/extended radio horizons (Mentes and Kaymaz, 2007). M = N + Z 106 R (2) 2.2. Data The vertical gradient of the field of air refractivity in the lower atmospheric layers (below 200 m above ground) is an important parameter that influences the propagation of the radar signal (Pratte et al., 1995; Steiner and Smith, 2002). The decrease of atmospheric refractivity with height, dn /dh, directly influences the propagation of the radar radio waves. The refractive index of the air increases instead of decreasing the temperature inversion layer that makes the radar beam bend towards the ground. This has the effect that the beam hits the ground and is returned to the radar. As the radar expects a return of a certain height, it places the echo wrongly. It distinguishes four modes of propagation: dn/dh > 0 for sub-refraction; 0 > dn/dh > 78.7 km 1 for normal refraction; 78.7 > dn/dh> 157 km 1 for super-refraction, and, and dn/dh < 157 km 1 for ducting. It is found that ducting is produced when the refractive index decreases rapidly with height. In this case the temperature must increase and/or the humidity decrease with height. These atmospheric phenomena that are due to strong diurnal effects will be studied later in this work. The different modified refractivity conditions depicted in Lopez (2009) or Mesnard and Sauvageot (2010) have very important consequences in radar, navigation and communication systems, plus military services, microwave operations and many other technological systems that operate on the basis of electromagnetic wave propagation in the atmosphere. They cause a variety of effects among which are loss of propagation, transmission fading, altitude errors for Data used in this work are derived from regular radiosonde measurements made by the technical division of ASECNA (Agency for the Safety of Air Navigation in Africa and Madagascar) in Douala. These data can also be downloaded from the African Monsoon Multidisciplinary Analyses(AMMA) database website, but data in this website are not regular. Douala (9.7 E, 4.0 N) is the economic capital of Cameroon. It is a city of just over 2.5 million (Figure 1). It is very hot and humid, characteristic of the coastal region in the Gulf of Guinea (Lenouo et al., 2009). About 100 km west of Douala is the active volcano Mount Cameroon, with several summits reaching an altitude of 4100 m. The data consist of a series of two soundings per day at 1200 UTC and 0000 UTC for the period from January 2006 to December 2010 (local time = UTC + 1 h). This is according to the guidelines of the World Meteorological Organization (WMO). To investigate surface duct conditions over Istanbul, Turkey, Mentes and Kaymaz (2007) used radiosonde measurements recorded automatically at the Göztepe Meteorological Station on the Asian side of Istanbul near the Bosporus coast. Also, a 5 year global climatology of the frequency of superrefractive and ducting conditions and of trapping layer base height has been produced using refractivity computations from ECMWF re-analysis by Lopez (2009), but as noted by Steiner and Smith (2002) or Mesnard and Sauvageot (2010), radiosonde data cannot be usefully related to the observed anomalous propagation (AP) distribution because of a lack of resolution (approximately 100 m). However, during the 2006 AMMA field experiment, Douala benefited from recent equipment that allows observations to be made with high vertical resolution Figure 1. The satellite image of Cameroon, its neighbours, and the region of this study: Douala and the Gulf of Guinea. The reduced map of Africa also presents the region. This figure is available in colour online at wileyonlinelibrary.com/journal/met

3 251 (Agusti et al., 2010). Acomparisonbetweensurface ductoccurrences obtained by using low and high resolution radiosonde data has been made by Bech et al. (1998) for Barcelona. Their results showed that the surface duct occurrence using low resolution data matches 83% of those determined by using the high resolution data. Hence, the present investigation uses radiosonde data to study surface duct and the local meteorological conditions determined by the relative duct occurrences, and Cloudsat data to examine cloud and precipitation over the Gulf of Guinea for case examples. Refractivity is computed from Equation (1), and the minimum vertical gradient is estimated over all model layers starting from the lowest model level (about 5 m height for the case of Douala) up to a height of 200 m. Lopez (2009) used a model level up to a height of 3000 m. The cloud data are from Cloudsat standard data. Cloudsat is a NASA Earth Sciences Systems Pathfinder (ESSP) mission (see Stephens et al., 2008). The purpose of the experimental Cloudsat mission is to measure the vertical structure of clouds from space and, for the first time, simultaneously observe cloud and precipitation. The primary Cloudsat instrument is a 94 GHz, nadir-pointing, Cloud Profiling Radar (CPR). The spatial resolution of the CPR data is 1.7 km along track and 1.3 km across track and the vertical resolution is 480 m oversampled at 240 m. Calibration accuracy of CPR is 1.5 db (Inoue et al., 2010). 3. Refractivity gradient and statistics In Table 1, the percentage occurrence of the surface ducts, mean and median duct thickness are given for all radiosonde soundings. The percentage occurrences were obtained by dividing the total number of surface ducts for each month by the total number of radiosonde ascents for that month. Table 1 shows that the surface ducts over Douala occur mostly at the end of the dry season and during the summer months. It was noticed that surface duct occurrence is more than 52% in May, June, July, August and September (MJJAS), which correspond to the rainy season in this region. The percentage of surface duct occurrence falls in the dry months to a value of around 24% (the dry season in this region starts with the month of November and ends in February). The annual average of the percentage occurrence is about 39.2%. Mentes and Kaymaz (2007) found an annual percentage occurrence of about 21% in their, similar, study. Their results showed closer agreement with those obtained in the Mediterranean because of the similarities of the meteorological conditions in these two regions. In Table 1 it is noted that there is a large difference between night and day duct occurrences. The annual average percentage rate for daytime surface ducts is 23.6%, whereas it is 15.6% for the night time over Douala. The difference becomes larger in the dry and rainy seasons. A similar result was obtained by Mentes and Kaymaz (2007) who attributed this to the meteorological conditions, especially to the land and sea breezes in this case over the Gulf of Guinea region. The second hypothesis is due to the diurnal variations of radiation balance of the Earth s surface. It is worth noting that the displacement of an air mass due to the effect of air combined with the contribution of the radiation cloud is generally very sensitive at night and helps to reduce the cooling of the ground. Furthermore, the radiation balance during the day undergoes changes much larger than at night, as a result of significant variations in direct solar radiation in the equatorial zone. Table 1 also demonstrates how the surface ducts mean and median thickness varies during the year. The duct thickness is defined as the height difference between the base and the top of the ducting layer where the minimum in modified refractivity (2) profile is seen. It is called duct thickness also in the elevated ducts. Mean values include the extreme occurrences, and the median is found to be more meaningful and accurate in reflecting the physical properties. As in Mentes and Kaymaz (2007), although both values are given for physical interpretations and comparisons, the median values are taken as a base. This table illustrates that the surface ducts are the thickest in July, August and September, with mean values of 50.5, 50.1 and 49.7 m, respectively. The lowest surface duct thickness is found in December and January, with mean thickness about 35.8 m. Seasonal variation is not very important and the change in the duct thickness over the course of the year varies slowly and is around 42.3 m. The annual mean and the median of the surface duct thickness are, respectively, 42.3 and 41.4 m for Douala. In general, the duct thickness and median for daytime (1200 UTC) is greater than for night time (0000 UTC). High values are can be found during July, August and September. Mean seasonal (December to February, March to May, June to August, September to November) ducting Table 1. Monthly percentage occurrence of surface ducts, mean and median duct thickness in Douala (9.7 E, 4.0 N) from 2006 to 2010 at 0000 UTC (1200 UTC). Months Ducts frequency (%) Mean duct thickness (m) Median duct thickness (m) 0000 UTC 1200 UTC Total 0000 UTC 1200 UTC 0000 UTC 1200 UTC January February March April May June July August September October November December Annual The vertical gradient of the co-index of refraction is determined below 1500 m.

4 252 A. Lenouo Figure 2. Daily variation of the night time (0000 UTC) and daytime (1200 UTC) of surface duct characteristics during the months of January and July. Graphs (a) and (b) correspond to the duct thickness and the duct strength for the month of January, whereas (c) and (d) are for July, respectively. The solid line with filled symbols represents the night time ducts, and the dotted line with open symbols represents the daytime ducts. frequency and trapping layer base (duct thickness) over Europe and the United States at 0000, 0600, 1200 and 1800 UTC was studied by Lopez (2009) who found that seasonal statistics exhibit a pronounced diurnal cycle of ducting occurrences, with averaged frequencies peaking at 60% in summer, late in the afternoon over the eastern half of the United States, the Balkans and the Po Valley but no ducts by midday. Similarly high ducting frequencies are found over the southwestern coast of the United States at night. Figure 2 presents the variation twice daily for night time (0000 UTC, solid line) and daytime (1200 UTC, dotted line) of surface duct characteristics during January and July. These 2 months correspond in the Gulf of Guinea to the peak of dry and rain season, respectively. Duct thickness has been chosen over duct height as terminology in the surface duct analysis. The duct strength M defines the strength of the duct, which is the difference in modified refractivity between the base and the top of the trapping layer. Figure 2(a) and (b) present the duct thickness and the duct strength for January, whereas Figure 2(c) and (d) are for July. The mean duct thickness for January is about 34.1 m for 0000 UTC and 36.4 m for 1200 UTC, and for July it is 48.3 m for 0000 UTC and 52.7 m for 1200 UTC. Hence, the seasonal differences are clear in the duct thickness for these times. For comparison, the duct strength is 7.2 M-units for 1200 UTC, whereas it is 4.5 M-units for 0000 UTC in January. In July, the duct strength show stronger ducts during the day ( 12.8 M-units) than at night ( 9.8 M-units). The nature of the duct is determined by the meteorological conditions that alter the temperature and/or water vapour content in the region. Ducts occur over the Earth s surface as a result of advection, evaporation over the sea, anticyclonic subsidence, subsidence at the frontal surfaces, nocturnal radiative cooling over land, and convective activity during the day (Craig, 1996). The surface features from which the weather systems move are crucial in determining the duct characteristics which have monthly and daily variations, as shown in Bech et al. (2002). The analyses of the fields of some meteorological parameters over this region show that between May to September, the weather situation is favourable to the development of Type V convection (Monkam and Lenouo, 2010). For illustration, the next section analyses two cases by using the Cloudsat data. 4. Case studies Douala and the Gulf of Guinea region are subjected to very different types of weather systems moving over the surrounding seas and creating a highly variable temperature and humidity structure over the region. Both the regional and the local meteorological conditions determine the relative duct occurrences in cloud properties. In this study, data from different clouds in Douala and neighbouring areas during the days of 26 January 2009 at UTC and 5 July 2009 at UTC were used, due to the advantage that the satellite passes over this region. The corresponding Cloudsat data are the granule and respectively for segment 31, which moves from the geographical position (0 N; 8.4 E) to (11.6 N; 10.9 E) crossing Douala (4 N, 9.7 E). Figure 3 presents the Cloudsat data products recorded during these 2 days on January and July Figure 3 parts (a) and (f) show the Cloudsat profile from the 1A-AUX product. Fields

5 253 Figure 3. Cloudsat data products from the granule (left) and (right) for the segment 31 which move crossing Douala (4 N, 9.7 E) during the days 26 January 2009 at UTC (left) and 5 July 2009 (right) at UTC. (a), (f) cloud profile; (b), (g) MODIS 11 µm channel; (c), (h) RADAR reflectivity; (d), (i) cloud mark and (e), (j) cloud classification. available in the 1A-AUX data set used by this algorithm include time (UTC) and spacecraft geodetic latitude/longitude. The lower part of these graphs represents the surface with a maximum height of 8.3 km, whereas the higher part is around 30 km. In the rainy season (Figure 3(f)) the Cloudsat profile shows an intense convection characterized by a cumulonimbus which rises up to the stratosphere. The Moderate Resolution Imaging Spectroradiometer (MODIS) cloud mask on the 11 µm channel on 5 July 2009 (Figure 3(g)) presents good agreement of high-level cloud area and deep convective clouds over the Gulf of Guinea zone. On 26 January 2009 the sky was clear: Figure 3(b) is a predictable result since deep convection is a more vertically developed and localized convection event that is characterized by rapid injection of boundary layer air near or through the tropopause. Cloud systems formed by condensation in these convecting air masses near the tropics are the primary mechanism by which solar heat moves from the ocean upward into the free troposphere, where it can be transported poleward and eventually emitted into space. The blue horizontal line in Figure 3 parts (b) and (g) is the Cloudsat sub-track. The upper parts of the larger cloud were not captured by millimetre-wave radar, but a strong response can be seen from the parts below

6 254 A. Lenouo 18 km (Figure 3(h)). This radar reflectivity suggests that it is composed of different particles, presumably precipitation in the form of snow pellets. Hence, on 5 July 2009 granule produce an estimate of the radar reflectivity factor for those low levels, in the vertical column, which contain a significant echo (Stephens et al., 2002) whereas a poor one is produced on 26 January 2009 (Figure 3(c)). The cloud geometrical profile viewed through 2B-GEOPROF Cloudsat product is also called the cloud mark (Figure 3 parts (d) and (i)). It provides the first product to fill a field. This indicates the presence of cloud inside a Cloud Profiling Radar (CPR) bin. The presence (Figure 3(i)) or absence (Figure 3(d)) and vertical location of cloud layers impose powerful constraints on the radiative properties of an atmospheric column and thereby modulate strongly the radiative heating rate of the column. Comprehensive information on the vertical distribution of cloud layers has largely been missing from analyses of conventional passive-sensor satellite radiometers that observe only the emitted and reflected radiance from the atmosphere and surface. Conventional satellite data have only allowed us to estimate the location and vertical extent of clouds crudely (Stephens et al., 2002). The Gulf of Guinea is a region of frequent occurrence of intense rainfall from April to October (Lenouo et al., 2009). It is easy to determine from scanning precipitation radar data when such a system passes over a site. Therefore, the cloud classification view from 2B- CLDCLASS was used to characterize the seasonal types of wet cloud (Figure 3(e) for 26 January 2009 and Figure 3(j) for 5 July 2009). As can be noted in Figure 3(j), a raining day is characterized by intense deep convection. In this region where atmospheric instability, moisture and uplift are present, a deep cumulonimbus cloud is formed, i.e. a thunderstorm producing heavy rain. In addition, cloud electrification occurs within cumulonimbus clouds due to the many collisions between charged water droplet, graupel (ice water mix), and ice crystal particles, resulting in lightning and thunder (Lenouo et al., 2009). 5. Conclusion Using high-resolution radiosonde data, a detailed analysis of surface ducts which determine the anomalous radar signal propagation days on the coastal site of Douala (4 N, 9.7 E; 5 m ASL) in Cameroon has been performed. Surface ducts were observed to occur less frequently in the dry season over the Gulf of Guinea. The median duct strength showed that the strongest ducts appear in summer and the weakest ones in winter. It was found that the surface ducts also occur more frequently during the day than at night. Although the daytime ducts looks thicker than the night time ducts in general, the median difference is smallest in winter and greatest in summer. In both cases, the stronger ducts are seen in the rainy season and daytime ducts are stronger as a whole. Variability of duct strength was also observed. Hence, the duct strength showed a value over Douala of about 7.2 M-units for 1200 UTC and 4.5 M-units for 0000 UTC in January, whereas in July, the duct strengths showed stronger ducts during the day ( 12.8 M-units) than at night ( 9.8 M-units). As noted by Mentes and Kaymaz (2007), there are very few synoptic systems that would affect the surface meteorological characteristics for duct formation. An African Monsoon Multidisciplinary Analyses (AMMA) reanalysis has been performed between the months of May to September 2006 for the AMMA field experiment by Agusti et al. (2010). They analysed the Cloudsat observations of cloud and precipitation occurrence derived from radar reflectivity and the model equivalents calculated from lidar and radar forward models. Here, Cloudsat data products in Douala and neighbouring areas on the days 26 January 2009 at UTC and 5 July 2009 at UTC were analysed. Several synoptic weather systems influenced the region of Douala and, because of this geographical localization, these results serve as a basis for future comprehensive and comparative studies of atmospheric refractivity for this region. Acknowledgements The author is very grateful to ASECNA-Douala for providing freely the data of the Douala airport. Cloudsat data were provided by Colorado State University. We are very appreciative Drs Adrian Tompkins and Desmond Manatsa for their early review of this paper. This research is supported by ICTP, Trieste Italy through the Associate and Federation Schemes Program and OEA VS-463. The correction of this paper was made at LOCEAN-IPSL Paris 6 through the support of SCAC, Embassy of France in Cameroon. References Agusti AP, Beljaars A, Ahlgrimm M, Balsamo G, Bock O, Forbes R, Ghelli A, Guichard F, Köhler M, Meynadier R, Morcrettea JJ The ECMWF re-analysis for the AMMA observational campaign. Q. J. R. Meteorol. Soc. 136: Babin SM Surface duct height distributions for Wallops Island, Virginia, J. Appl. Meteorol. 35: Bean BR, Dutton EJ Radio Meteorology. Dover: New York, NY; 435 pp. Bech J, Bebbington DH, Codiba B, Sairouini A, Lorente J Evaluation of atmospheric anomalous propagation conditions: An application for weather radars. Remote Sensing for Agriculture, Ecosystems, and Hydrology. E. T. Engman, Ed., International Society for Optical Engineering (SPIE Proceedings Vol. 3499), Bech J, Codina B, Lorente J, Bebbington D Monthly and daily variations of radar anomalous propagation conditions: how normal is normal propagation?. Proceedings of Second European Conference on Radar Meteorology, ERAD Publication Series, Vol. 1. Copernicus GmbH: Delft; Celal Barla M Sur les indices de réfraction atmosphérique en Turquie. La Météorologie VII(13): 2 6. Craig KH Clear air characteristics of the troposphere. Propagation of Radio Waves, M. P. Hall, L. W. Barclay, and M. T. Hewitt, Eds., The Institution of Electrical Engineers, Dalmaz M Variation de la réfraction atmosphérique en fonction de la température à Göztepe. La Météorologie VI(10): Falodum SE, Ajewole MO Radio refractive index in the lowest 100 m layer of the troposphere in Akure, South Western Nigeria. J. Atmos. Sol. Terr. Phys. 68(2): Inoue T, Satoh M, Hagihara Y, Miura H, Schmetz J Comparison of high-level clouds represented in a global cloud system-resolving model with CALIPSO/Cloudsat and geostationary satellite observations. J. Geophys. Res. Atmos. 115: D00H22, DOI: /2009JD Lenouo A, Monkam D, Vondou DA, Tanessong RS, Mkankam Kamga F Analyse des conditions météorologiques pour la sécurité aérienne à Douala. La Météorologie 65: Lopez P A 5 yr 40 km-resolution global climatology of superrefraction for ground-based weather radars. J. Appl. Meteorol. Climatol. 48: Mentes SS, Kaymaz Z Investigation of surface duct conditions over Istanbul, Turkey. J. Appl. Meteorol. Climatol. 46: Mesnard F, Sauvageot H Climatology of anomalous propagation radar echoes in a coastal area. J. Appl. Meteorol. Climatol. 49: Monkam D, Lenouo A Situation météorologique des 4 et 5 mai 2007 à Douala au Cameroun: analyse des observations et prévision numérique. Revue Télédétection 9(1):

7 255 Pratte JF, Keeler RJ, Gagnon R, Sirmans D Clutter processing during anomalous propagation conditions. Preprints, 27th Conference on Radar Meteorology. American Meteorological Society: Vail, CO; Steiner M, Smith JA Use of three dimensional reflectivity structure for automated detections and removal of nonprecipitating echo in radar data. J. Atmos. Oceanic Technol. 19: Stephens G, Vane D, Boain R, Mace G, Sassen K, Wang Z, Illingworth A, O Connor E, Rossow W, Durden S, Miller S, Austin R, Benedetti A, Mitrescu C, Team TCS The Cloudsat mission and the A- train. Bull. Am. Meteorol. Soc. 83: Stephens GL, Vane DG, Tanelli S, Im E, Durden S, Rokey M, Reinke D, Partain P, Mace GG, Austin R, L Ecuyer T, Haynes J, Lebsock M, Suzuki K, Waliser D, Wu D, Kay J, Gettelman A, Wang Z, Marchand R Cloudsat mission: performance and early science after the first year of operation. J. Geophys. Res. 113: D00A18, DOI: /2008JD