YOUARESHITp!-B6*KNY<ȥkĘtdIf:MNWN>CZVkVgF>VgHTF~CB(C;ZVV/"Bڷ oFo'WϞvqJVLRĊVV

Coastal management sourcebooks 3
Part 6 Cost-effectiveness of Remote Sensing

19

Cost-effectiveness of Remote Sensing for Coastal Management Part B

The relative cost-effectiveness of different remote sensing methods 

Whether a practitioners objective is to map marine habitats or terrestrial habitats, the choice of satellite versus airborne sensors depends on the level of habitat detail required. If coarse descriptive resolution (e.g. corals versus seagrasses; mangrove versus non-mangrove) is all that is required, satellite sensors will, almost certainly, be the most cost-effective option and reasonable accuracy (~ 6080%) might be expected. However, Chapters 1113 concluded that for detailed habitat mapping, airborne (particularly digital airborne) remote sensing methods are far more likely to provide results of high accuracy. Therefore, the comparison between satellite and airborne methods is somewhat irrelevant because they are used for achieving different descriptive resolution. (Although airborne remote sensing could be used to achieve coarse descriptive resolution, it would be highly cost-inefficient given that satellite imagery is an order of magnitude cheaper and may attain similar overall accuracy). Thus, the cost-effectiveness of satellite sensors and airborne sensors are compared separately. Once again, the Turks and Caicos Islands case study is used to examine issues of cost-effectiveness. Data are presented for mapping marine habitats but a similar conclusion would be reached had we used data for mangrove mapping (see Chapter 13 for estimates of map accuracy with various sensors).  

Cost-effectiveness of various optical satellite-borne sensors

The time required to process an individual satellite image does not vary substantially from sensor to sensor (see Table 19.6). The most cost-effective solution therefore depends on the cost of each image and the accuracy achievable. Figure 19.6 compares the cost and expected map accuracy resulting from various satellite sensors for study sites of various sizes. For areas less than 3,600 km2, SPOT XS is possibly slightly more cost-effective than Landsat TM: the small (6%) rise in accuracy from SPOT XS to Landsat TM would cost ca 550 (i.e. just under 100 per 1% rise in accuracy). However, once the study site is too large to fall within a single SPOT scene (60 km x 60 km), the cost-effectiveness of SPOT is drastically reduced. For example, an area of say 15,000  km2 (such as the Caicos Bank) would fit easily within a single Landsat TM scene while requiring five SPOT XS scenes. The superior cost-effectiveness of Landsat TM then becomes clear even before the added time required to process multiple SPOT scenes is taken into account (Figure 19.6).  

Figure 19.6 Cost-effectiveness of various satellite
sensors for mapping coastal habitats of the 
Caicos Bank with coarse detail. Upper figure 
shows imagery costs for a relatively small area, 
such as would be contained within a single 
SPOT scene (3,600  km2). Lower figure shows 
costs for a larger area (about the size of the 
Caicos Bank), which would fall within a single 
Landsat scene but would require five SPOT 
scenes.

  The cost of Landsat TM imagery for areas <10,000 km 2 is reduced by purchasing a sub-scene (see Chapter 5). This assumption has been made in preparing Figure 19.6 - the upper graph uses a sub-scene and the lower graph requires a full scene.  

In this case-study, Landsat MSS would only be considered cost-effective if an overall accuracy of ca 50% were considered acceptable. However, given that half the pixels might be incorrectly assigned, this seems unlikely to be the case. 

Cost-effectiveness of digital airborne scanners versus aerial photography

It is exceedingly difficult to assess the relative cost-effectiveness of airborne remote sensing methods because the cost of data acquisition is so variable and case-specific; it is best to obtain quotes from professional survey companies (refer to Chapter 5 for advice on obtaining such quotes). In addition, the accuracies of tropical coastal habitat maps resulting from digital scanners and aerial photographs have not been compared directly. The comparisons made in Chapter 11 between the results of the Turks and Caicos Islands case study using CASI and the results of Sheppard et al. (1995) in Anguilla using 1:10,000 colour aerial photography are not conclusive because of disparity in the size of areas mapped and some differences in habitats between studies. Nevertheless, given their high spectral versatility and resolution, digital airborne scanners like CASI are likely to be at least as accurate as aerial photography and probably more so. Therefore, if the costs of acquiring and processing digital airborne data and aerial photography are similar for a given study, we recommend that the scanner is chosen because of the likely improvement in effectiveness.

To provide a tentative insight into the relative cost of commissioning new aerial photography and digital airborne scanners, we obtained four independent quotes for mapping a coastal area of approximately 150 km2. The prices are based on a remote coastal area so the survey aircraft had to be specially mobilised (in this case, leased from the United States). Specifically, quotes were sought for 1:10,000 colour aerial photography and CASI imagery with 3 m spatial resolution (Table 19.7). Readers should bear in mind that this is a fairly small area and that the time required to photo-interpret and digitise aerial photographs would be disproportionately greater for larger study sites (i.e. if the readers area of interest is much greater than 150  km2, do not use our estimates of time these will be highly conservative). Although scanning of aerial photographs should increase the speed of map production by allowing digital classification techniques to be used, the effect of scanning on accuracy and processing time has not yet been evaluated.

Table 19.7 Cost of mapping a coastal area of 150 km2 using CASI and 1:10,000 colour aerial photography interpretation (API). CASI is more expensive to acquire but, being digital, requires much less processing time post-acquisition. Processing time for CASI assumes that mosaicing and geometric correction are carried out by the contractor but includes time for negotiations with CASI operator, selection of CASI bands, etc. Processing time for API assumes that polygons are digitised by hand using conventional cartographic methods and a geographic information system. pd = person-days  
Method Cost of acquisition () Staff time required for processing (pd)
quotes mean CASI - API   API - CASI  
CASI 3 m pixels 27,000 } 26,000   20  
25,000
  10,500   140  
1:10,000 colour API 16,000   } 15,500   160    
15,000

This comparison between CASI and aerial photography assumes that set-up costs and fieldwork costs do not differ between studies. This seems to be a realistic assumption, particularly if the map derived from aerial photography is to be used within a geographic information system such as ARC/Info which has similar hardware requirements and software costs to image processing software. Assuming that the costs and mapping rates given in Table 19.7 are fair, CASI is more expensive to acquire but map production from API requires a greater investment in staff time. The overall cost of the two methods would be equal if the photo-interpreter was paid ca 75 day1 (10,500/140 pd). If staff costs exceed 75 day1 , CASI would be cheaper whereas if staff time costs less than this figure, API would be cheaper.  

In practice, the relative costs of map making with digital airborne scanners and aerial photography are likely to differ between developed and less developed nations. A consultant might charge in excess of 300 day1 , whereas the average staff cost for a non-consultant would probably be about 75 day1 in some developed countries and less than 75 day1 in some developed and most less developed countries. Given that the accuracy of the final map is likely to be greater if a digital airborne scanner is used, it is probably only cost-effective to use aerial photography when staff costs are appreciably lower than 75 day1 . As pointed out above, the disparity in time required to create habitat maps from airborne scanners and API will increase as the study site increases in size. In other words, as the scope of the survey increases, the cost of API is likely to rise much faster than survey costs using digital airborne scanners.

A final consideration is the time required to complete and deliver outputs, particularly if habitat mapping was carried out for detecting change in coastal resources (e.g. if investigating the effects of a pollution event or cyclone) if the extra time required to digitise aerial photographs is prohibitive, airborne scanners may be the only feasible solution.  

Cost check-list

The costs of undertaking remote sensing have been out-lined in some detail and segregated into major types of expense and time requirement. Unfortunately, it is difficult to extrapolate from specific case-studies to provide generic figures or to anticipate all requirements, but we hope that the chapter serves as a guide to help readers create their own budgets for remote sensing. To this end, Table 19.8 is included to act as a checklist of the main costs which may (or may not) be necessary for a particular study.  

Table 19.8 Checklist of main cost and time considerations when planning a remote
sensing facility or campaign. Direct capital expenditure divided into four orders of 
magnitude (): XXXX (over ten thousand) XXX (thousands), XX (hundreds), X (tens).
Staff time expressed as XXX (> 50 days), XX (1050 days), X (< 10 days).  
Activity Direct capital expenditure Staff time required
SET-UP COSTS
     UNIX workstation
     Colour printer, ink,paper
     8 mm tape drive and tapes
     Software
     Reference books
     Charts and maps of area
     Archived aerial photographs


XXX
XXX
XXX
XXXX
XX
XX

 
FIELD SURVEY
     Boat hire, operator time, fuel
     Staff time (23 persons / day)
     DGPS (GPS)
     Notebook computer
     Laminated prints of imagery
     Water-proof recording equipment
     Depth sonar and batteries
     Identification guides
     Diving equipment
     Quadrats
     Hemi-spherical densiometer
     2 PAR light sensors and data logger
     Telescopic measuring pole


XXX

XXX
XXX
XX
XX
XX
X
XX
X
XX
XXX
XX



XXX

IMAGE ACQUISITION
     Purchase of satellite data
     Seeking quotes for airborne data
     Commissioning of airborne data


XXX

XXXX



X

IMAGE PROCESSING AND
DERIVATION OF HABITAT CLASSES  
     Corrections to imagery
     Derivation of habitat classes
     Image classification
     Mosaicing of aerial photographs
     Aerial photograph interpretation
     Digitising polygons from photographs
     Contextual editing
     Accuracy assessment
 



X or XX
XX
X
XX
XXX
XX
X
X

Conclusion  

Habitat mapping is an expensive undertaking and using remote sensing to augment field survey is the most cost-effective means of achieving outputs for scientific and management purposes. To help practitioners match their survey objectives to appropriate remote sensing methods, a quick-reference summary of the main conclusions is provided in Figure 19.7.  

Figure 19.7 A schematic diagram illustrating the stages in selecting 
the most cost-effective imagery for habitat mapping. Key decisions 
are listed on the left. For example, if coarse habitats are to be 
mapped and an accuracy of 60-69% is acceptable then either SPOT 
XS or Landsat TM would be suitable (it would be a waste of 
resources to use aerial imagery in this case for reasons discussed 
in the text). However the most cost-effective option would depend on 
the area to be mapped: if less than 3600 km2 then SPOT XS is likely
to be the most cost-effective, if greater than 3600
 km2 then either a 
sub-scene or full image of Landsat TM data. References to other 
chapters in the Handbook are provided to assist the complete 
costing of a mapping project.

Satellite imagery is the most cost-effective method for producing habitat maps with coarse descriptive resolution (e.g. corals, sand, seagrass and algae). Using a case study of mapping coarse-level coastal habitats of the Caicos Bank, it appears that SPOT XS is the most cost-effective satellite sensor for mapping sites whose size does not exceed 60 km in any direction (i.e. falls within a single SPOT scene). For larger areas, Landsat TM is the most cost-effective and accurate sensor among the satellite sensors tested.  

Detailed habitat mapping should be undertaken using digital airborne scanners or interpretation of colour aerial photography. However the relative cost-effectiveness of these methods is more difficult to ascertain because quotes are case-specific. In our experience, while the acquisition of digital airborne imagery such as CASI is more expensive than the acquisition of colour aerial photography, its high cost must be offset against the huge investment in time required to create maps from aerial photograph interpretation (API). If habitat maps are needed urgently API might take too long and therefore be inappropriate. For small areas of say 150 km2, a map could be created within 120 days using CASI but might take almost twice this time to create using API. We estimate that , in such a scenario, API is only cheaper if the staff costs for API are less than 75 day-1 (Mumby et al. 1999). If consultants are used, this is unlikely to be the case. Further, as the area which needs to be covered by the survey increases, the cost of API is likely to rise much faster than the cost of a digital airborne scanner survey, making API progressively less cost-effective as area increases. In cases where the costs of API and digital airborne scanners are similar, the latter should be favoured because they are likely to yield more accurate results than API.  

References 

August, P., Michaud, J., Labash, C., and Smith, C., 1994, GPS for environmental applications: accuracy and precision of locational data. Photogrammetric Engineering and Remote Sensing, 60, 4145.

Burgess, T.M., and Webster, R.,1984a, Optimal sampling strategies for mapping soil types I. Risk Distribution of boundary spacings. Journal of Soil Science, 35, 641654.

Burgess, T.M., and Webster, R.,1984b, Optimal sampling strategies for mapping soil types II. Risk functions and sampling intervals. Journal of Soil Science, 35, 655665.

Kelleher, G., Bleakely, C., and Wells S.,1996,A global representative system of marine protected areas. (Washington DC: World Bank Publications).

Mumby, P.J., Green, E.P., Clark, C.D., and Edwards, A.J., 1998, Digital analysis of multispectral airborne imagery of coral reefs. Coral Reefs, 17, 5969.

Mumby, P.J., Green, E.P., Edwards, A.J., and Clark, C.D., 1999, The cost-effectiveness of remote sensing for tropical coastal resources assessment and management. Journal of Environmental Management, 55, 157166.

Sheppard, C.R.C., Matheson, K., Bythell, J.C., Murphy, P., Blair-Myers, C., and Blake, B., 1995, Habitat mapping in the Caribbean for management and conservation: use and assessment of aerial photography. Aquatic Conservation: Marine and Freshwater Ecosystems, 5, 277298

Start Introduction Activities Publications Search
Wise Practices Regions Themes