Impact Evaluation of HIMA Iringa Region Tanzania

Size: px

Start display at page:

Download "Impact Evaluation of HIMA Iringa Region Tanzania"

Adam Pierce
5 years ago
Views:

Ringstedvej 20 DK-4000 Roskilde Denmark October 2007 Goss Gilroy Inc.

1 Ministry of Foreign Affairs (Danida) Impact Evaluation of HIMA Iringa Region Tanzania Annex 7: Satellite Image Analysis Prepared jointly by: Orbicon A/S Ringstedvej 20 DK-4000 Roskilde Denmark October 2007 Goss Gilroy Inc. Management Consultants Suite 900, 150 Metcalfe Street Ottawa, Ontario K2P 1P1, Canada

2 1.0 Concepts of Using Satellite Images as a Monitoring Tool In the ToR defining activities for the Evaluation, it was anticipated that satellite images were to be used in the HIMA implementation period and that the images should be used for before and after assessment of physical HIMA interventions. However, satellite images were not used in any documents despite several reports stating that such analysis could be valuable for monitoring purposes. Although considered several times by various project management teams as well as the RDE staff, the HIMA project never purchased any satellite images. In 2000, air photos were taken of divisions in both Makete and Njombe districts. The photos from Makete were missing and a series of photos in a scale of 1:50,000 covering Imalinyi and Lupembe divisions in Njombe division were available. Unidentified air photos from 1985 and 1986 (incomplete series) from the former IRADEP (Iringa agricultural development project, EEC) in a scale of 1:25,000 were also found. The incomplete regional IRADEP photos from 1985 and the HIMA photos from 2000 are missing the flight run maps, are lacking in even basic initial ground interpretation, as well as field verification and do not include photos over several years within the impact measurement period. Ultimately, the photos are of no value to the Evaluation. A Danish funded research project on sustainable agriculture using a combination of satellite images and aerial photos was carried out in the period in two geographical confined areas (approximately 60 km 2 ) of Ikuwala Village (Mazombe Division) and Ilambilole Village (Isimani Divison), both in Iringa Rural District. However, the two villages were not HIMA villages and the end point of the fieldbased research was Thus, the images cannot give information pertinent to the HIMA intervention. However, the underlying reasons in land use changes and livelihood are well documented 1. Planned attempts to undertake change detection analysis of vegetation and land use as part of the impact evaluation were therefore not possible from existing HIMA monitoring activities. To mitigate the implications for the impact evaluation the Evaluation suggested purchasing images targeted to the issue. An overview of the main options considered in utilising satellite images is summarized in the Table 5.1 and consulted with Danida. 1 The research program was named SASA - Sustainable Agriculture in Semi-arid Africa. The findings from Ikuwala and Ilambilole village are reported in: A livelihood Perspective on Natural Resource Management and Environmental Change in Semiarid Tanzania by Torben Birch-Thomsen, et.al. Economic Geography, Vol 77, No.1, January An interview with T. B. Thomsen as key informant has been made on 9. February 2007 in Copenhagen by Steen Dahl and Birte Hansen 2 page 2 of 9

3 Type of Satellite Images Resolution Images available before + after HIMA period ( ) QuickBird 0.6 m no (in operation only >2002) Ikonos 1 m no (in operation only >1999) LANDSAT or SPOT Relevance to impact evaluation Highly relevant at village level (ability to determine village details: houses, fences, woodlots, gullies, etc.) Highly relevant at village level (ability to determine village details: houses, fences, woodlots, gullies, etc.). ~20-30 m yes Relevant at catchments level (when determining general changes detection in land use patterns). Not useful at village level. Technical details and limitations a) Archive Images (min. 5 x 5 km 2 ) can be provided within few days. Only minor parts of the Iringa Rural and Mufindi are covered. b) New Images (min 8 x 8 km 2 ) can be provided within few weeks if the area is cloud free. a) Archive Images (min. 7 x 7 km 2 ) can be provided within few days. Only parts of Iringa Rural are covered. b) New Images (min 10 x 10 km 2 ) can be provided within few weeks if the area is cloud free. Archive LandSat images can be provided before 95 and after 02 and changes in main categories of land cover can be identified and sized in sub-catchments of say km 2. Interpretation will be significantly improved if supplemented with Ikonos or Quickbird images. Table 1.1 Overview of various types of satellite images, relevance to impact evaluation and technical details. The coverage of the Quickbird archive images in Iringa Region is scattered and only covers the southern part of the Iringa Rural and a few images in Mufindi. The Ikonos archive has coherent series in the northern part of Iringa Rural District. The option of purchasing high-resolution images to be used on village level was selected as first priority. The rationale was that the diversified and heterogenic HIMA intervention was more adequately documented using high-resolution images rather the LANDSAT/SPOT images with low resolution. The Ikonos or Quickbird images have proved very useful tools in obtaining, i) an overview of the village/sub-villages and locations, ii) as a crankshaft for discussion in focus groups on changed practice in agriculture and natural resources management, and iii) as a base map for orientation when specific HIMA activities are to be verified in the field. Furthermore, laminated prints of the satellite images were left behind in the individual villages and are expected to be a useful tool for the 3 page 3 of 9

4 Location name Tarafani, Isimani District Itimbo Iringa District Mdasi Njombe District Image Njombe District Mang oto Makete District villages in future planning. Five villages were selected for further assessment using images (Table 5.2). However, cloudy conditions in November and December 2006 prevented recording of new images of three villages. The Evaluation therefore used archive images (2005) of Tarafani and Itimbo in the further assessment. Latitude south Longitude east Image type Ikonos archive (49 km 2 recorded on 9. September 05) Ikonos archive (49 km 2 recorded on 9. September 05) QuickBird - new image (64 km 2 ) QuickBird - new image (64 km 2 ) QuickBird - new image (64 km 2 ) Table 1.2 Locations selected for satellite image assessment. Recordings from Mdasi, Image, and Mang oto in Novemver-December 2006 were cancelled due to cloudy conditions. The interpretations of images from the two available archive images villages are highlighted in the subsequent sections. 2.0 Image Interpretation of the Tarafani Village Area, Isimani Division 2.1 Settings and Area Description Tarafani is located in the Isimani Division in Iringa District (former Iringa Rural District). The village has a population of 1625 inhabitants and 402 households. The households are concentrated along the Iringa Dodoma gravel road, approximately 50 km north of Iringa. The area is in the dry part of the Iringa Region in the transition between AEZ 8 and 16 having harsh agricultural conditions. Tarafani is situated on a gentle NE sloping terrain with minor depressions along two river courses being tributaries to Great Ruaha River. The rivers form the northern and southern part of the village boundary. Tarafani became part of the HIMA programme in 1998 taking over from CONCERN. CONCERN is an NGO that had been active for more than a decade in Tarafani with activities similar to HIMA. HIMA took over the premises of CONCERN and the buildings were refurbished to become a HIMA sub-centre. The 4 page 4 of 9

5 buildings are still in use (2006) and well maintained. However, activities were carried out in the last few years of the HIMA program only. According to information from the villagers, the HIMA interventions in Tarafani were limited to training in conservation of woodlots and forested areas, improved bee keeping and attempts to protect a water point source next to the river south of Tarafani. From the HIMA sub-centre, support was allocated to the district and divisional staff in training and supervising activities and thus not necessarily linked to the Tarafani village as such. The villagers claimed that most of the HIMA support was allocated to the HIMA staff activities outside the village. Due to the similar activities of CONCERN and HIMA distinguishing the impacts of the two programs is difficult. Villagers seemed to be more focused on the results achieved by CONCERN activities due to their much longer involvement. By contrast, the HIMA activities seemed short and abrupt and did not fulfill the expectations of the villagers. 2.2 Site Verification of HIMA Activities Using Satellite Images A site inspection in Tarafani was carried out on the 7 th of December Six villagers including the village chairman formed a group with the overall objective of pinpointing visible interventions of HIMA. The areas of interests were identified on the high resolution image (7 x 7 km 2 Ikonos image recorded on the 9 th of September 2005 with 1 m pixel resolution) and subsequently inspected together with the group. The approximately location of the village boundary has been redrawn on the image based on information retrieved from a handmade sketch (resource map) seen in the village office (ofisi ya kijiji). The village boundary on the image cannot be considered official nor very accurate. 5 page 5 of 9

6 6 page 6 of 9

7 3.0 Image Interpretation of the Itimbo Village Area, Kilolo Division 3.1 Settings and Area Description Itimbo Village is located in the Kilolo Division and is part of Kilolo District (former Iringa Rural District). The village has a population of 3004 inhabitants and 441 households. The village is spread into seven hamlets. The distance to Iringa is 40 km by gravel road. The area is within the favorable precipitation belt of AEZ 3, but close to the drier AEZ 8 and 14. The terrain is dominated by soil covered rolling hills exhibiting altitude differences of roughly 50 m to 150 m. Some of the hillsides have fairly steep gradients and small streams are frequent within the valleys. The valley floors are usually under intensive cultivation. In the vicinity of the hamlets, most hillsides are cultivated or covered with woodlots. The hills further from the households are not cultivated but have grass and some trees. The landscape has many ravines and gullies. HIMA activities were introduced in Itimbo under Phase II ( ). The activities included agro-forestry (tree nurseries, planting instructions, new timber wood and fruit trees varieties), contour ridging (and introduction of Vetiver / Guatamala grasses), and protection of water sources. Further details on the HIMA activities can be retrieved from the PRA report. 3.2 Site Verification of HIMA Activities Using Satellite Images A site inspection in Itimbo was carried out on the 8 th of December Six villagers including a HIMA para-professional formed a group with the overall objective of pinpointing visible interventions of HIMA. The areas of interests were identified on the high-resolution image (7 x 7 km 2 Ikonos image recorded on the 9 th of September 2005 with 1 m pixel resolution) and selected plots subsequently inspected together with the group. The villagers immediately recognized the details and orientation of the images and were able to precisely locate major physical intervention areas (woodlots, water point protection sites, contour bund areas). Consequently, the group prepared the area demarcations seen on the image in joint effort with the Evaluation. Following a common discussion and consensus in the group, the villagers demarcated the village boundary on the image. The demarcation was done of the major fix points along the village boundary (hill tops, confluences of rivers, etc.). 7 page 7 of 9

8 8 page 8 of 9

9 4.0 Concluding Remarks on the Use of Images in Rural Impact Evaluations The use of high-resolution satellite images (the Ikonos or QuickBird type) facilitated identification of details in the village area including individual houses and plots. This was useful when HIMA activities were to be discussed and verified at village level as the HIMA interventions were highly diverse and heterogeneous. The use of satellite images from two villages in Iringa demonstrated that villagers could easily recognize and orient themselves and discuss the physical impacts in the landscape. The villagers perceptions can be applied to a broader context and can help determine the extent of impact. These images in the PRA part of the evaluation could supplement the hand drawn resource maps. 9 page 9 of 9

DEVELOPMENT OF A NEW SOUTH AFRICAN LAND-COVER DATASET USING AUTOMATED MAPPING TECHINQUES. Mark Thompson 1

DEVELOPMENT OF A NEW SOUTH AFRICAN LAND-COVER DATASET USING AUTOMATED MAPPING TECHINQUES. Mark Thompson 1 1 GeoTerraImage Pty Ltd, Pretoria, South Africa Abstract This talk will discuss the development