Chapter V: Meteorology

99.	About 50 posts were set up during the course of the Survey, of which 20 were maintained for the whole seven years, January 1944 to December 1950, or longer. Of the 25 posts manned by General Survey observers, 15 completed the seven years observation (some from October 1943, to early January 1951).
100.	Some General Survey posts were closed down during the Survey for reasons of economy or difficulty of supervision. Some of the R.A.F. stations and police and Illao posts were also closed down, and the official closure of licensed coffee-shops resulted in difficulty in maintaining some posts. The R.A.F. posts at Berbera and Hargeisa were collecting very much more data than has been attempted by the General Survey. These two posts were maintained by the R.A.F. for most of the seven-years Survey.
101.	The Senior Met. Officer, R.A.F., Aden, and the Chief Met. Officer, Nairobi (at first R.A.F.), were most helpful with advice, records and some equipment. The most valuable series of records obtainable in the area, however, is that maintained since 1900 by the Observatory staff at Jibuti in French Somaliland. Records there are kept partly by trained Somali observers. The Chef du Service Météorologique has been most helpful and the Observatory is to be congratulated on having maintained continuous records for so long despite world wars, political changes, and often conditions of extreme discomfort.
102.	The General Survey observers were recruited from an otherwise untapped reserve of literate Arabic writers: some of these were ex-police writers, freed by the increasing use of English in the Police Department. Mr. Hassan Nur, Senior Nation Assistant Surveyor, translated most of the Arabic reports into English, and quickly learned to use the Gazetteer (Table 3, para. 79), to not latitudes and longitudes of place names on his translations. The plotting of the translated and gazetted reports was carried out exclusively by Viney and Hunt, who were sufficiently informed about the Protectorate as a whole to check the gazetted reports whilst plotting the meteorological and tribal data onto monthly work-maps.
103.	Employment was not found for most of the Arabic writers at the end of the Survey, but they are a ready source of clerks and observers for any department which employs an assistant capable of translating the local "bush Arabic' widely used in the Protectorate.
104.	As the patrolling of posts, partly by motor vehicle and partly on foot, took about six weeks to complete, it was soon found necessary to reduce the number of reports to one at the end of each month, when a patrol with rations and pay visited the post and collect the report, noting other data during the patrol. The number of mislaid reports was thus reduced to a minimum.
105.	Equipment was obtained only with difficulty in the early states of the Survey (1943 - 44), but as soon as a few thermometers were obtained maximum and minimum temperatures and humidities were recorded at as many posts as the thermometers would serve. Ground wind was also observed at some posts, and direction of could movement at a few, but the first concern of all observers was always the accurate recording of rainfall at 8.30 A.m. every morning, including nil records, and the reporting of rainfall in the neighbourhood of the post. The records so obtained vary in reliability, but by constant monthly patrolling by observer sergeants, native assistant surveyors or survey officer, and occasional transfers, replacements on leave, and odd checks by travelling officers of other departments, a fair average of reliability seems to have been obtained. Outstanding records have always been follow up, enquiries made from officials who were in the area and, if necessary, personal visits made to interrogate local Somalis, and to note any effects of flood, the state of grazing, etc. Such outstanding rain records were, e.g. those at Bihen-Nogaled (Post 18), 14 .55 inches in May 1945; and at Daloh (Post 25), 19.00 inches in June 1949.
106.	The degree of accuracy obtained for the final average rainfalls is a function not only of the number and position of observer posts, and the efficiency with which observers reported on rainfall between their posts, but also of the period of time during which observations were carried out. Rainfall is known to be extremely sporadic both in time and areally, and it is believed that unusually large falls of rain do in fact occur sporadically across the country in most years. With only 30 to 50 rain gauges such cloudbursts are seldom recorded, but are known to occur, and are particularly noticeable in the areas of low average rainfall of the E. and S.E.

B. General Climatic Scheme (See Somaliland Contour Sketch Map, illus. 6, para. 72)

107.	The Survey was concerned primarily with the Protectorate, and adjacent areas grazed by British Protected tribes, and the neighbouring countries of French Somaliland, Ethiopia and Somalia (Italiana) have not been thoroughly covered..
108.	Roughly the area of highest rainfall (10 - 20 inches) is the area over 4,000 feet above sea level (see illus. 6) consisting of the Harar Plateau in the west, the Golis, Wogr, and Ashararet Ranges in the central Protectorate, and the Al Hills of the northe-east. This area is the Main Watershed of the country. These areas get some rain in most months, and do particularly well in the period between the main rains of April to June and the short rains of October-November, when many areas are drying up in the desiccating Kharif (Haga), S.W. Monsoon wind.
109.	South of this plateau belt the rains tend to fll mostly in April to June and October to November, any other minor rains falling mostly on the ribs of land extending south-westward from the main plateau (see ills. 5, para. 71, 3,000-foot contour).
110.	North of the Main Watershed, and on the east coast of Somalia, there are more frequently sporadic rains during the dry season from December to March, presumably due to the upward deflection of the N.E. Monsoon by the Main Watershed mountains, the steep scarps of which face the Gulf of Aden.
111.	The climate appears to depend firstly on the fact that the sun passes vertically overhead twice in the year, with the resultant N.E. Monsoon when the sun is in the south, and S.W. Monsoon when it is north of the Protectorate. The hot season is from April to September, when the sun is north of the Equator.
112.	Between the end of the N.E. Monsoon and the beginning of the W.W. Monsoon is a calm, windless period in April called "Kalil" in Somali. In April the main "Gu" rains should begin (but see Table 2, para. 58, Somali Seasonal Calendar), in the south and west: in the south because the season is naturally earlier in the south whence the sun has come, and in the west because of the altitude of the Harar Plateau. This Gu rain spreads to the north and east during April (sometimes delayed until May). In a good year it generally rains every day (most often in the afternoon) for three or four days running, followed by a period of two or three days without rain, during April and the first half o May. By June the rain has usually become less, and the dry S.W. Monsoon definitely unpleasant. In July there is usually not very much rain, and south-west gales are not unusual. The highest temperatures of the year are recorded in July and August though owing to the wind the heat is less oppressive than in the Kalil calm periods of April and September.
113.	Usually in the second half of July rain starts again at the higher altitudes and continues along the Main Watershed ("Kalarug" in Somali), until the clam Kalis of September, when the S.W. Monsoon drops. In october the N.E. Monsoon period should start, and there is nearly always widespread rain (though not so much as in the Gu main rains). In November there should be of this "Dhair" rain in the first half of the month. By November, howerver, the N.E. Monsoon should be blowing quite strongly, and in an average year there is little rain in the second half of November, or in December and January, though in some years the scarps of the Main Watershed facing the Gulf of Aden get quite good rains in these months.
114.	In February there are usually rains in the west and in isolated parts of the Watershed, very often in the mornings (Maie). In March there is usually increasing rain in the west (Harar Plateau), and sometime over the whole Watershed area and the high ribs of land. In the west this rain often runs right on into the main Gu rains of April again.

115.	The assessment of the value of rainfall for a year depends not only on the number of inches recorded at posts, the number of posts, the area covered, and the amount of information between posts, but also on the amount of rain which falls in a day and is able to penetrate the soil before being evaporated or running off.
116.	In the apparently porous red soil of the Haud, over which water had trickled after heavy thunderstorms on a nearly level surface for 24 hours (Qaidr Boleh, 1949), the water penetrated only 9 inches. Much research remains to be done on porosity of soils, run-off, evaporation, etc. That the water from the Main Watershed does penetrate the soil is proved by the existence of the belt of wells in the line Hargeisa, Guled Haji, Odweina, Burao (and the Ain and Nogal valleys). In the Hargeisa-Burao well zone there is plenty of water, and that at Burao, 80 to 90 feet below the surface, certainly comes from the Golis-Wogr watershed rains.
117.	The day-to-day records of rain from posts have been kept only in manuscript form (packed away in cases in Hargeisa).
118.	Despite the actual statistical records (Tables 5 and 6, paras. 139, 140), and the average rainfall maps (illus. 14 - 28, paras. 124 - 138), the assessment of values of rainfall (Table 8, para. 142) has been made largely from the annual and quarterly maps, and the monthly work-maps. The last are only in manuscript form, but annual and quarterly rainfall maps have been published in Annual Reports, 1944 - 49, and the distribution factor in assessment of rainfall values is derived from them.
119.	It has been found that isohyets (lines of equal rainfall, cf. contours showing lines of equal altitude), need not necessarily completely embrace each other as must altitude contours. To climb from 100 to 300 feet one must pass the 200-foot contour; but there may be a cloudbursts on one side of a line and no rain at all on the other. Isohyets have, therefore been drawn without adherence to the usual practice of making the lines embrace, but attempting to show the actual facts. In some countries isohyets will in face resemble contours. In the course of a long period, even in areas of sporadic rainfall like Somaliland, the average maps of rainfall will in the end probably resemble contour maps; but for shorter periods one must expect the ishoyets to butt up against each other, in a way which no draughtsman of contours would approve.
120.	Annual rainfalls such as Go'o 43.68 inches in 1946, Daloh 41.11 inches in 1945, and especially Bihen (Nogaled) 18.89 inches in 1945, have been doubted. When, however, one finds the old record of Sheikh 47.14 inches in 1923, and experiences a storm in the Golis, or the Al Hills, one is convinced that such records are quite likely. The terrific storm in May 1945 at Bihen in the usually dry S.E. Nogal, was fortunately observed by Captain Gilliland who was making a botanical survey there, though he fled before the storm and did not actually read the gauge. In 1941, however, the writer's transport was marooned for ten days to a half-mile perimeter by floods at Gardo, and simultaneously Messrs. Smith, Brook, and Harris-Rivett were similarly held up in different parts of the Nogal, whilst the Tug Der at Burao overflowed its banks, flooding and destroying some of the Government buildings. There is no doubt that astounding falls of rain do occur in the Protectorate, and though unusual records should always be checked, the few high records of the seven-year General Survey probably show a fair proportion of the storms which occur either in small areas, or more rarely, over most the Protectorate.
121.	For several years there have been three rain-gauges at Hargeisa (R.A.F. on landing ground at 4,500 feet, and District Commissioner's and Agricultural Office's at 4,2000 feet, half a mile apart and five miles north of the R.A.F. Station). There has also sometimes been an extra gauge at the Police Office in Burao, about a quarter of a mile from the General Survey Headquarters gauge. On the whole these gauges near each other give similar readings over a year (though not identical), but frequently one records a storm on a day when there is no rain at all at the post nearby.
122.	Table 5 (below) and the Summary of Table 5 in Table 6, show the actual recorded rainfall statistics. The maps (illus. 14 - 28 below) are the quintessence of about 130 monthly, quarterly and annual rainfall maps (mostly on the scale 1 : 3,000.000). The maximum maps, both quarterly and annual, have been compiled by taking the highest rainfall plotted for any point on nay of the seven annual report maps: the minimum by similarly taking the lowest plotted records. The resulting compilation maps have then been checked with the statistics given in Table 5, and slightly amended where necessary.
123.	The average rainfall maps have been compiled by taking pairs of maps, quarterly or annual, for the seven-year period thus: 1944 - 1945 1946 - 1947 1948 - 1949 1944 - 1947 1948 - 1950 1944 - 1950 and the final map 1944 - 50 checked and amended from Table 5 as with the maximum and minimum maps. It should be noted that the maximum and minimum maps show the highest and lowest rainfalls recorded in seven years, not for the Protectorate as a whole, but of the extremes in different places for different years in a patchwork. The average rainfall maps include all the results of the whole seven-year survey.
124. - 138.	(Illustrations 14 - 28.)

Notes for Writers: Bookmarks are included for each section, i.e. "A." - "H.", and will later be included for each numbered paragraph, i.e. "443." to "555.", each table and each illustration.

96.	Whilst data concerning many sciences can be collected at any time, it is essential that meteorological recording should be continuous over as wide an area and for as long a time as possible. The same applies to the collection of data concerning tribal movements (see Chapter IX). Observations posts were therefore set up to cover as much of the country as possible.
97.	At first it had been hoped that existing Government posts and licensed coffee shops could be used, and that officers of other departments could supervise many of the posts. In fact the whole administration of the country was in a state of flux, and it was found necessary to man and patrol independent posts, to ensure that the work was properly supervised, the observers rationed and paid regularly, and that records reached Survey Headquarters.

98.	(Illustration 13.)