The community-based participatory mapping represents a useful tool for urban mosquito larval control. After its completion, corrected sketch maps, description forms and formalized colour maps based on an aerial photograph were available for the complete study area. On this basis, 100% spatial coverage of mosquito larval habitat surveillance by CORPs was achieved, which would have been impossible with either the sketch maps or the formalized colour maps alone.
From the point of view of programme field workers including CORPs, the sketch maps and associated detailed plot descriptions are indispensable guidance tools. The sketch map system accommodates the different cognitive abilities of the CORPs, as the map style can be adapted according to their personal preferences in order to achieve optimal orientation. However, only few CORPs were comfortable to use an aerial photograph as a basis for their work, which rules out the option of replacing all sketch maps with formalized maps. Nevertheless, CORPs who wish to use formalized maps as an addition to their sketch maps can be provided with laminated printouts. When a CORP has to be replaced, the successor takes over the existing sketch maps but is free to adjust or redraw them if desired.
From a programme management perspective, the sketch maps are an ideal method to assign a unique number to each plot, whereas the technical mapping approach with aerial imagery proved to be essential for the verification and correction of the sketch maps. Moreover, the georeferenced colour maps that show the demarcations and locations of TCUs enable management staff to assess and analyze the data collected by the CORPs, and to conduct targeted spot checks.
The use of GIS software in the mapping approach proved to be extremely helpful for programme management and supervision of field activities, although only basic functions were utilized. Similar positive findings have also been made in other malaria control programmes in South Africa and Mozambique [34–36], and public health in Africa generally . The approach does not require any electronic devices such as GPS receivers in the field. In addition, if digital aerial imagery is available, costly equipment like digitizing tablets or large format scanners are not needed. The entire GIS database as well as all subsequent updates thereof has been made available to the central GIS unit of the Dar es Salaam City Council. It can be used as a basis for any Council activity such as health interventions, waste management programs, and urban planning, to name a few.
The mapping approach adheres to the existing administrative boundary system in Tanzania, mainly referring to the ten-cell-units. In a dynamic environment such as the rapidly growing city of Dar es Salaam, this allows optimal orientation for community-based programme staff in the field, without having to create entirely new sets of artificial boundaries. Whenever there are changes, sketch maps can easily be updated during their weekly use by the CORPs. The technical team only needs to be informed in case TCU boundaries have been modified. It is argued that this approach has practical programmatic advantages over imposed raster grid systems [23, 38, 39], because it considers user-definable boundaries that can be agreed in a participatory manner on the ground and that can be readily recognized by community-based staff without access to, or the necessary education to use, GIS technology. In this way, GIS can be participatory, with the potential to enhance community involvement . In the operational context of malaria control Dar es Salaam, this rather basic but straightforward way of applying GIS is advantageous, as resources in terms of available data and expert personnel are limited. The same tendency has also been observed for lower-income countries in general , and accessing such limited resources can be a challenge in itself.
The system of ten-cell-units such as the one in Dar es Salaam (or hamlets and vitongoji, as they are called in the rural districts of Tanzania ) probably is slightly different to the administrative systems in countries other than Tanzania. Therefore, applying this mapping approach to other regions of Africa and beyond will require the adaptation to the particular systems of each country. In such cases, the smallest administrative units that exist in the respective areas of interest in those countries can be used as adequate substitutes for ten-cell-units. However, for successfully utilizing the participatory mapping procedure, it is of crucial importance that the residents of the target areas are aware of the administrative units they live in. Otherwise, community-based programme staff would not be able to draw from the knowledge of community members regarding the locations of boundaries. Therefore, in areas where the smallest existing administrative units are not well known to the local population, it might be a good alternative to refer to roads, rivers, pathways or similar intuitive landmarks that can be easily identified by community members.
Similar mapping approaches in African settings have been implemented for other purposes, scales and cities. For example, in Southern Sudan, urban maps have been produced to assist town planners in their efforts to respond effectively to returning population and reintegration issues. The Data Exchange Platform for the Horn of Africa (DEPHA) also provides a few datasets on urban scale. The EPIDEMIO programme has produced maps of several African cities. However, in contrast to the procedure in Dar es Salaam, these GIS-based approaches required a considerable amount of technical expertise and external support. Moreover, there are no participatory components. Hence, they cannot provide the necessary basis for community-based comprehensive mosquito larval surveillance.
The costs of the mapping approach are listed in table 2. For mapping the entire study area surveyed here, less than US$ 14,000 have been spent, which is equal to US$ 831 per km2. Thus, the complete set of correct sketch and formal maps covering one TCU costs an average of approximately US$ 24. Considering that the maps have to be produced only once and do not require much updating from then on, these costs appear reasonable and affordable not only for the Dar es Salaam programme, but also for any other comparable larval control intervention in Africa or elsewhere.
Areas that were initially not included in any sketch map are theoretically just as likely as any other area to contain breeding sites for malaria vectors, and might be very important sources for mosquitoes that fly into residential areas. In the study area in Dar es Salaam, most of the newly included areas were industrial or commercial areas and open spaces. Whereas industrial and commercial areas might be just as important as residential ones in terms of mosquito productivity, the open grass and scrublands that often frame such industrial and commercial plots are particularly likely to support key vectors from the An. gambiae complex . This is particularly true for open spaces, notably those that are located in lowlands with a relatively high ground water table, and used for agricultural purposes. Considering that the number of infective mosquito bites per person per year is inversely proportional to the human population density [44, 45], and mosquitoes disperse until they find blood [46, 47], all these predominantly unoccupied areas might therefore contribute considerably to mosquito emergence rate [23, 38, 46–48] and malaria transmission  in neighbouring residential TCUs. Therefore, the inclusion of the initially unsurveyed areas into routine mosquito larval surveillance and control is likely to have a great impact on the effectiveness of such a programme, particularly after the planned addition of surrounding wards to the UMCP.
The framework generated through this mapping procedure made it possible to rationally allocate every square meter of the programme area to individual CORPs under the oversight of specific supervisors. Such individualization of responsibility is considered essential for managing larviciding programmes [50, 51] because of the rigorous, sustained and comprehensive coverage required to achieve useful reductions of malaria transmission in Africa.