CHANGE DETECTION 
Interest in change focusses on the impact of the dams. The Bakolori dam was completed in 
1978, since then the flow of the Sokoto River has been controlled. Selection of imagery for 
1975 and 1984 thus straddles a period in which there has been marked change. Initial work has 
concentrated on the middle Sokoto valley, including the area of the Bakolori scheme. Obviously, 
there is an extensive area of designed change within the confines of the irrigation scheme. There 
is change also downstream of this, where there is no longer the natural wet season flooding of 
the valley, which was a prominent feature of the natural regime, and where dry season flow is 
now sustained by release of water from the dam. 
TABLE 5. CHANGE RESULTS, BAKOLOR!, 1975 — 1984 
Phenoiogica: Human-induced 
Change . Change 
Losses Gains 
Land Use/Cover Category Pixels % Total  % Total % Total 
Pixels Pixels 
Wetland Non-vegetated 91 0.56 2.54 0.71: 
Wetland Vegetated : 483 2.99 4.34 430 
Fadama = 1153 517,13 17.93 5.63 
Surface Irrigation " Qm m 0.00 8.55 
Rice Cultivation * 0 _ 0.00 16.69 
: Sugar Cane * 0. _ 0.34... 0.16 
Rain-fed Cultivation (wet) 2802 12.33 17.22 : 22.40 i 
Rain-fed Cultivation (dry) 54 0.33 6.92 : 7.27 
Shrubland/Grass 8011 49.56 28.33 17.22 
Shrubland/Thicket - 2839 17.56 15.76 . 9.73 
Sandy/Hardpan 46. 0.28 5.07 1.55 
Laterite Slopes 25 0.16 0.19 : 0.15 
Burned Areas * 0 0.98 0.00 
Water Bodies 552 3.42 0.14 5.23 
Settlements 110 0.68 0.16 0.41 
Total 16,166 100.00 100.00 # 100.00 # 
* Identified only on one date, so no change detectable 
# Total pixels 54,334 in each column 
Adams (1986) has shown the strong relationship between duration of inundation and distri- 
bution of crops in the fadama, and from this it is clear that any modification of the natural flood 
regime will have profound impact on land use in areas outside the irrigation scheme. Effects are 
experienced in upland areas also, particularly where farmers who were displaced by the flooding 
of the reservoir have been resettled. Analysis of landscape change between 1975 and 1984 invol- 
ved an ‘‘enhanced classification’’ procedure developed by Pilon (1986). This procedure permitted 
the separation of phenological and human-induced change. Results (Table 5) show that change 
occured in just over 35% of the total study area (70,500 pixels out of 199,056).