1.4 Congo

In the Congo, the first phase of a pure conventional water investigation program was successfully completed by GTZ. The continuation of this project was to take place in the beginning of 1987 some 100 km further north-west near Loubomo, in an area hardly developed and essentially unexplored. Within the program small water supplies for scattered villages should be installed. Because the wells were to be operated with hand pumps, the required minimum yields were again only 5 l/min.

Success rates of 75% were reported for the first phase. The underground was composed of karst and was so cavernous that no kind of measuring technique was necessary for the location of well sites. As the second area was estimated to show similar conditions, a hydrogeologist decided very roughly locations for 50 possible well sites, solely on account of morphologic indications and purposefulness. In succession to that, Schröter was requested to determine the exact positions of 50 final drilling points by means of the dowsing technique. His duty consisted in indicating the supposed karst water resources as precisely as possible. Along these lines, Schröter indicated before the beginning of the drilling at each location the depth and yield that he expected.

Two years later, results of 51 borehole drillings were available; two of them, however, were not executed at positions determined by Schröter. One borehole collapsed at a depth of 39 m without having a water strike before the anticipated final depth of 50 m had been reached. Another drilling struck semi-artesian water at a depth of 52 m; the water rose in the borehole and clogged it due to an expanding clay underground so that this principally successful drilling had to be abandoned for the time being.

For 47 wells, the actually obtained data may be compared with the predictions. For 42 of them, the expected minimum yield had been accomplished. This represents a success rate of 90%. It should be added, that the actual yields were several times higher than the necessary minimum quantity and the maximum of the deliverable quantities would have been still higher; these maximum values have not been determined (there was no need to know them) and are expected to fluctuate with the season. It is therefore not possible to check the predicted yields in an absolutely meaningful manner. Moreover, since the indicated values vary only slightly from borehole to borehole, their significance may not be too high anyway. It remains to say that for half of the cases, the predicted values were accurate, whereby it should not be forgotten that the predictions had all been made before any drilling results were available in this area. An analogy with the former project is not very helpful because more clay components were found in the present project area.

Some comments are in order with respect to the 5 unsuccessful borehole drillings. Two of these holes had to be drilled up to a depth of 50 and 55 m, but the depths actually reached were only 47 and 50 m, respectively. As the predicted depths may only be interpreted as a guide line, a certain continuation of the drilling would have been advisable. Finally, a satisfactory evaluation of the result would have required answers concerning several other questions such as whether the unsuccessful boreholes had been drilled vertically and technically correct and whether clay intrusions may have impeded the water flow into the borehole.

In any case, the recipe "to drill deeper" is no guarantee for success. The three other unsuccessful drillings of respective depth of 50, 60 and 60 m have been considerably deepened to 78, 72 and 87 m, respectively, without getting any sign of water.

The predicted drilling depths in that partly hilly region with height differences of up to 50 m turned out to be practical and useful values. The distribution of the differences between expected and realized depths for 42 successful borehole drillings are compiled in Table 3, sorted into 4 groups. According to the data, 67% of the predicted depths were better than 10% accurate and 80% were correct within 20%. In the evaluation of the results, however, additional aspects have to be considered.

TABLE 3 Deviation Number Percentage 0 - 10% 28 67 % 10 - 20% 5 12% 20 - 40% 7 16% > 40% 2 5%

Table 3. Results referring to drilling depths of 42 wells in a karst area in the Congo. The drilled depths varied between 70 meters with a mean value of 48 m, the predictions varied between 40 and 60 meters with a mean value of 47 m. Listed are the absolute and proportional number of depths predicted by the GTZ expert, Hans Schröter, in different accuracy intervals. For example, 67% of the data were more accurate than 10%.

First, the mean value of the actual depths amounted to 48 m and about 80% of the drilling depths ranged from 40 to 60 m. Thus, a very rough and rather unspecific "prediction" of, say, 50 m depth for each drilling would automatically have resulted in a success rate of 80%, when an accuracy of ± 20% was aimed at. Indeed, the drilling depths advised by Schröter did range from 40 to 60 m with an average close to 47 m. They showed, however, a significant correlation with the actually required depths. Besides that, by resorting to "intelligent guessing" a mean drilling depth of 70 m should have been chosen; that number resulted from the experience obtained during the first project phase and the hydrogeologist in charge recommended it also as the basis for the new project planning regarding the expected total depth to be drilled. The quite different prediction for a mean depth of only 47 m could therefore (before the first drillings had taken place in the actual project area) not be well founded. Hints due to landmarks in conjunction with topographic features cannot be fully excluded, but must be regarded as highly improbable.

In general, boreholes were drilled only as deep as necessary in order to reach the required, relatively low minimum yield. Therefore, it cannot be argued that, after having found water, the boreholes had been drilled still deeper up to the initially advised depth. This was unnecessary for a further reason: in most cases, the relatively small yields aimed at were immediately found in karst water channels and thus, it was not necessary to carry on the drilling in order to improve the yield. However, a less significant trend in this direction can not be completely excluded because of self-interest of the drilling company which is usually paid for drilled meters irrespective of the results. Finally, it must be stated that, for the moment, no simple classical measuring procedures are available which permit detailed depth predictions for relatively small water channels in deep lying karst. In the meantime, 40 drilling spots were located by means of the hydrogeological method and, in part, by dowsing techniques during the third phase of the project in the geologically different Kiwongo District. No kind of drilling had been formerly carried out in this area. Up to now, 33 of the 40 points have been drilled, 19 of which are successful. It turned out that clay layers of some 40 m thickness cover a series of fractured calcareous sandstones. The water found in the fracture zones was semi-artesian and rose up to 10 - 15 m below the ground level. With 10 l/min per meter lowered level of the water table, the yields were clearly above the expected minimum values. The uniformity of the drilling profile indicated that a regionally extended aquifer exists. Since drilling points were not systematically divided according to their origins, i.e. whether determined by hydrogeological means or dowsing techniques, and as some boreholes are still to be drilled, the respective success rates cannot be ascertained.