Table 10.1a Information Gaps Identified for each Habitat Type, Gaps Filled by Field Surveys and Gaps Yet to be Filled |
| Habitat Type |
Information Used for Mapping and Possible Data Sources for Ecological Ranking |
Information Gaps Identified |
Information Gap Filled by Field Surveys |
Information Gaps Yet to be Filled |
| Fung Shui Forest |
The habitat was recognised in unsupervised classification and refined in supervised classification through input of information from preliminary site visits. Point locations of 115 fung shui forests identified by Chu (1998) and boundaries for 17 of these habitats were mapped onto the habitat map. |
Among the 115 fung shui forests identified, 98 of them did not have information on their boundaries and ecological value. |
Surveys were conducted on 25 point locations of fung shui forest and boundaries of these forest patches were mapped onto the habitat map. Ecological value of these habitats was verified. |
There are still 73 identified fung shui forests where boundaries are uncertain and ecological value is yet to be confirmed. |
| Montane Forest |
The habitat was recognised in unsupervised classification and refined in supervised classification through input of information from preliminary site visits. Differentiated from lowland forest based on altitude (ie > 600 m above sea level). |
Studies on montane forests were few and only that of Zhuang and Corlett (1997) had provided descriptions of the forest features of Tai Mo Shan in the New Territories and Sunset Peak on Lantau Island. Despite the information in Zhuang and Corlett (1997), in common with other areas of suspected montane forest in Hong Kong, more information would be required before the habitat type can be verified and ecological value assigned. |
A total of 14 patches (sub-sites) of montane forest were surveyed and their habitat nature and ecological value verified. |
There are still some montane forest patches (about 2.2 ha) left unsurveyed under the present study. These habitats are mainly located on steep slopes or in ravine valleys where, as indicated by our surveyor, accessibility is very limited. Further efforts on surveying the remaining area is considered not necessary given its small size and very restricted accessibility. |
| Lowland Forest |
The habitat was recognised in unsupervised classification and refined in supervised classification through input of information from preliminary site visits. Differentiated from montane forest based on altitude (ie below 600 m above sea level). |
As the habitat is likely to be blended with other vegetation types, such as mixed shrubland, Baeckia shrubland and plantation, uncertainties existed in distinguishing the habitat. Although there were a few studies undertaken on lowland forests (eg Zhuang and Corlett (1997) and Kwok (1996)), information on specific sites remained scarce. |
A total of 106 patches (sub-sites) of lowland forest were surveyed and their habitat nature and ecological value verified. |
There are still about 508 hectares of lowland forest identified on the habitat map left unsurveyed and their habitat type and ecological value remain to be confirmed. |
| Mixed Shrubland |
The habitat was recognised in unsupervised classification and refined in supervised classification through input of information from preliminary site visits. |
As the habitat is likely to be blended with other vegetation habitats, such as lowland forest, Baeckia shrubland and shrubby grassland, uncertainties existed in distinguishing the habitat from other similar vegetation types. |
A total 149 patches (sub-sites) of mixed shrubland were surveyed and their habitat nature and ecological value verified. |
There are still about 14,169 hectares of mixed shrubland identified on the habitat map left unsurveyed and, therefore, their habitat type and ecological value are yet to be confirmed. |
Freshwater/
Brackish Wetlands |
The habitat was mapped principally during the supervised classification which utilised input from known wetland areas delineated by Dudgeon and Chan (1996).
In-depth surveys were conducted for the 28 freshwater, unmanaged wetlands identified in the AFCD Freshwater Wetland Study by Dudgeon and Chan (1996) and the classification devised by the authors for assessing the degree of conservation importance of the sites could be useful in justifying the ecological value of these sites. As advised by AFCD (memo dated 14 May 1999), wetland habitats > 5 hectares will be (have been) covered by the AFCD Wetland Compensation Study. Information on these habitats can be useful for justification of ecological ranking of wetlands. |
As the habitat is likely to be blended with other wet, or damp vegetation due to spectral similarities between these habitats, uncertainties existed in distinguishing the habitat from other similar habitat types.
While the AFCD Study on wetlands (Dudgeon and Chan 1996) could be useful for justifying ecological value of the 28 wetlands, the surveys were actually conducted some years ago and since wetlands were highly variable in nature and vulnerable to disturbance, more updated information would be required for final ranking of these habitats. Few data on other wetlands, such as those within Country Parks (ie managed) and those which were brackish in nature, were available to justify adjustment of the ecological value rating of high.
As advised by AFCD (memo dated 14 May 1999), wetland habitats > 5 hectares would be (had been) covered by the AFCD Wetland Compensation Study. Therefore, wetlands < 5 hectares would require field surveys to verify their habitat type and ecological rating. |
A total of 254 patches (sub-sites) of wetlands were surveyed and, wherever possible (as some are private land and inaccessible for surveying), their habitat nature and ecological value were verified. |
About 285 hectares of wetlands were surveyed by the SUSDEV 21 Study Team. There are still about 887 hectares of wetlands identified on the habitat map not surveyed under the present study, of which 713 hectares were not covered by the AFCD Study by Dudgeon and Chan (1996). The existing status and ecological value of these habitats remains to be confirmed. |
| Natural Watercourse |
The habitat was mapped by using 1:1K digital topographic data from the Lands Department. EPD (1998) River Water Quality in Hong Kong was used to justify the necessity to further adjust natural watercourses into the modified category based on their pollution status. |
Despite the detailed Lands Department data, some of the junction areas for modified and natural watercourses could not be clearly identified and thus required field surveys to check the classification of natural and modified watercourse boundaries. Data on the ecological status of local natural watercourses are largely from surveys conducted during the 70's or 80's which are not sufficiently up-to-date, or from studies which are species-oriented and thus not comprehensive enough to derive a finding on conservation value. Field surveys were therefore necessary to evaluate the conservation status of some of the habitats identified. |
A total 54 locations (sub-sites) of natural watercourse were surveyed and their habitat type and ecological value verified. |
Apart from those watercourses surveyed under the present study, there are still 578 hectares of natural watercourses yet to have their habitat type and ecological status confirmed. |
| Mangrove |
The habitat was mapped through supervised classification using existing information from Tam and Wong (1997) and aerial photos as the primary data sources. Mangroves have been extensively studied by Tam and Wong (1997). A total of 43 mangrove stands have been identified with 23 of them studied in detail. In the study's report, a conservation value was assigned to each of the mangrove habitats (relative to other mangrove habitats) based on the results obtained during field surveys. The designated value can be used as a reference for assessing whether the indicative ecological conservation value of high is appropriately applied to each site. |
There were habitat areas which have been identified under imagery classification but lack existing data to substantiate. Field surveys were thus required to verify the habitat status and ecological value of these habitats. While the AFCD Study on mangroves (Tam and Wong 1997) was useful for justifying ecological value of the 23 stands, the surveys were actually conducted some years ago and since mangrove habitats are vulnerable to disturbance, field surveys were therefore required to update the conservation value ranking of the mangrove stands. |
A total of 21 patches (sub-sites) of mangrove were surveyed including some of those which have been classified as having the lowest overall conservation scores by Tam and Wong (1997) and those which might have been affected by human disturbance. Existing status and boundary of these stands were verified and ecological value confirmed during site surveys. |
Only a small portion of the total habitat (about 1 ha) was covered under the present study. Apart from those where existing information is available to substantiate their existence and ecological rating, there are still some mangrove stands which lack verification. |
| Seagrass Bed |
As areas of seagrass beds are difficult to be distinguished through satellite imagery classification and aerial photos, they were mapped primarily based on existing information provided by the AFCD Herbarium and published materials on seagrass ecology (eg Fong (1998a, b), Fong et al (1998), SWIMS (1994) and Wu and Lee (1998)). |
Information on the extent and location of existing seagrass beds were considered sufficiently comprehensive and up-to-date. In terms of ecological value, seagrass beds were also considered to be well-studied. |
One-day effort was used to survey areas which are likely to support a larger seagrass population (including Sham Sha Po, So Lo Pun and Sam A Chung). |
While information on the extent and location of existing seagrass beds were sufficiently comprehensive and up-to-date, periodic checking on the identified seagrass beds may be required as some of the seagrass species are highly seasonal and might vary in areal extent from year to year (Fong, pers comm). |
| Intertidal Mudflat |
The habitat was mapped primarily by manual digitising of identified intertidal mudflat habitats compiled through reviews of 1:20K topographic data, aerial photos and specialist input. As the landward side of mudflats is usually fringed by mangroves, locations of mudflats were checked against the existing 43 mangrove stands in Hong Kong identified by Tam and Wong (1997). Besides mangroves, seagrass beds are often associated with mudflats as the latter provides substrate for seagrass growth. Identified locations for seagrass beds were used as a reference to locate mudflat areas. Those possible flat areas not indicated on the Hong Kong Map Series were cross-checked with the satellite imagery and aerial photos.
Detailed studies on the ecological status are few, except for some studies on larger flat areas such as Mai Po (eg McChesney (1997), Chan (1990) and Choi (1991)) and Shui Hau (Chiu and Morton 1999). Only a few mudflats in Hong Kong (including Pak Nai, San Tau and Shui Hau) are well-documented as a high ecological value habitat due to the presence of ecologically important species (horseshoe crabs) (Chiu and Morton 1999). |
Since it was not possible to identify all intertidal mudflat areas using the mapping method described, uncertainties existed in some of the mapped locations and required field surveys to validate their existence and boundaries. There were flat areas which were not distinguishable on the 1:20K Lands Department maps and there was no existing information available for locating their boundary. These mudflats required field surveys to validate their boundary.There were a number of mudflat habitats identified on the habitat map which lacked adequate information against which the conservation value could be justified. |
Boundary and ecological value of 24 mudflat locations (sub-sites) identified on the habitat map were verified through field visits. Mudflats which were small in size and likely to be affected by human disturbance (eg coastal development), such as Tai Ho Wan, were also included for field surveying. |
There are still some identified mudflat habitats left unsurveyed under the present study or lack sufficient existing information to substantiate their indicative high ecological ranking. Efforts may be required to verify the boundary and ecological value of these habitats. |
| Shrubby Grassland |
The habitat was recognised in unsupervised classification and refined in supervised classification through input of information from preliminary site visits. |
Although preliminary site visits were conducted in an attempt to provide input to improve the classification, uncertainty remains high for areas where the relative proportion of grass and woody plants are similar (eg potential intermingling with mixed shrubland habitat) even though classification has somewhat been improved. As there is little information on the ecological value of shrubby grassland habitats, individual habitats may receive a lower or higher ranking only if they are the subject of field surveys. |
A total of 64 patches (182 ha) of shrubby grassland were surveyed and their boundary and ecological value verified. |
There are still about 8,514 hectares of shrubby grassland not covered by the present baseline survey and their existing boundary and ecological value remain to be confirmed. |
| Baeckia Shrubland |
The habitat was recognised in unsupervised classification and refined in supervised classification through input of information from preliminary site visits. |
As Baeckia shrubs often mix profusely with other shrubs and trees, many areas and boundaries of Baeckia shrubland could not be mapped with confidence and required site surveys to check the classification and boundary of many areas. As there is no published information on the habitat value of Baeckia in Hong Kong, and only sporadic information on species which may use Baeckia habitats, field surveys were required before adjusting the ecological ranking of individual Baeckia sites. |
A total of 59 patches (175 ha) of Baeckia shrubland were surveyed and their boundary and ecological value verified. |
There are still about 5,826 hectares of Baeckia shrubland not covered by the present baseline survey and their existing boundary and ecological value remain to be confirmed. |
Plantation or
Plantation/
Mixed Forest |
The habitat was mapped through a visual identification of patterns in tree planting and through input of information from preliminary site visits. |
As the majority of plantations have become mixed plantation forests and intermingled with other spectrally and visually similar habitats (eg lowland forests), they are unable to be identified and accurately mapped using the satellite imagery and aerial photographs.The inclusion of "in recognizable rows from the air" in the definition has excluded substantial areas in Hong Kong where plantations are known to exist (eg areas near reservoirs) and the total area mapped under this category has therefore been under-estimated. |
More plantations were picked up by surveyors during site visits to lowland forest, mixed shrubland, wetland, shrubby grassland, Baeckia shrubland, cultivation and grassland. |
As the majority of the re-classified plantations are actually intermingled with lowland forest and mixed shrubland near reservoirs or within Country Park areas, efforts may be required to survey the lowland and mixed shrubland habitats not covered under the present baseline survey. |
Fishpond/
Gei Wai |
The habitat was mapped primarily through recourse to the existing data compiled under the Study on the Ecological Value of Fishponds in the Deep Bay Area supplied by the Planning Department, and the 1:20,000 topographic map from the Lands Department. The results of the mapping were then checked against the aerial photos. Information on the conservation status of fishponds and gei wais is available for the Mai Po Marshes (eg Melville and Morton (1983) and Irving and Morton (1988)). |
Numerous fishponds were surveyed under the PlanD study and therefore ecological data was available for these habitats. Fishponds and gei wais not surveyed under the Plan D study nor included in studies of the Mai Po area required field surveys to validate their existing status and ecological value. |
A total of 166 fishpond/gei wai habitats were surveyed and, wherever possible (as some are private property and inaccessible for surveying), their existing status and ecological value were verified. |
There are still some fishponds/gei wais where no surveys have been conducted and therefore lack existing information to substantiate their existence/boundary and ecological value. |
| Sandy Shore |
The habitat was mapped through recourse to manual identification of the habitats based on existing topographic data (Lands Department 1997) and the expertise of the Study Team. Aerial photos were cross-referenced as necessary. Recent ecological surveys have been conducted for sandy shores on the North Lantau coastline (Scott Wilson 2000), and previous studies have investigated the communities of some wave-exposed sandy beaches (eg Big Wave Bay) by Morton and Morton (1983) and Wong (1990) and such sheltered beaches as Tong Fuk Miu Wan on South Lantau (Morton and Morton 1983) and Tai Tam (Ong Che and Morton 1994). |
While the habitat was considered to be mapped with a high level of accuracy, some spot-checking with field surveys were required. There were still a number of sandy beaches not surveyed in the past which thus lacked information to substantiate their ecological status. Field surveys were required to supplement the information. |
A total of 9 beaches (sub-sites) were surveyed to verify their areal extent and confirm their ecological ranking. |
There are still some sandy beaches where no surveys have been undertaken and therefore no data are available to justify their ecological value. Further survey efforts may be devoted to those sheltered sandy shores which are located in very remote areas, eg North-east New Territories, where upgrading of indicative ecological value may be necessary (as is the case with the survey sites at Fung Wong Wat (Site 131) and Wong Chuk Kok Hoi (Site 132). |
| Rocky Shore |
The habitat was mapped through manual identification using existing 1:20K topographic data from Lands Department, aerial photos and the expertise of the Study Team. Many ecological surveys have been undertaken on rocky shores, including Cape d'Aguilar (Morton and Harper (1995) and Morton (1997)), Chung Hom Kok (Wadpole 1985), Green Island (Babtie 1999), Hoi Ha Wan (Liu 1992), Lamma Island (ERM 1999), Lo Chau, Po Toi and Soko Islands (ERM (unpublished data)) and North Lantau (Scott Wilson 2000). |
Mapping confidence is thus high with only a few of the coastline areas requiring field-checking.While many of the local rocky shores have been extensively studied, some other areas have never been studied and required field surveys to confirm their conservation status. |
Habitat type and ecological status of a total of 9 locations (sub-sites) of rocky shores were verified and justified. |
There are still some rocky beaches where no surveys have been undertaken and therefore no data are available to justify their ecological value. Further efforts may be required to survey those exposed rocky shores which are located in very remote areas, eg outlying islands, where upgrading of indicative ecological value may be necessary (as was the case with the survey sites at North Ninepin Island (Site 133) and Basalt Island (Site 134). |
| Cultivation |
Because of the heterogeneous nature of this habitat type and spectral similarities with other categories (eg grassland, shrubby grassland, wetland), the cultivation habitat type cannot be mapped accurately using imagery classification alone. Therefore, 1:20K topographic data from Lands Department have been used to identify the areas through manual discrimination of cultivated areas from spectrally similar but non-cultivated areas, and has substantially improved the mapping confidence. |
As the habitat is likely to be co-mingled with other vegetation habitats, such as shrubland and grassland, uncertainties exists in distinguishing the habitat from other similar vegetation types. Field surveys were required to clarify whether the area is actively under cultivation. There is little existing information on the ecological value of cultivated lands and thus ecological ranking would be impossible without field surveys. |
Habitat type and ecological rating of a total of 45 patches (sub-sites) of cultivation were confirmed and justified. |
There are still about 4,010 hectares of cultivation mapped on the habitat map left unsurveyed and uncertainties remain in these areas with regard to their habitat type and ecological status. As the habitat is highly heterogeneous in nature, further efforts may be required to differentiate cultivated land from other habitat type(s) to justify ecological rating. |
| Bare Rock or Soil |
The habitat was mapped through supervised classification and distinguished from other spectrally similar categories (including landfill, quarry, rocky shore and artificial rocky/hard shoreline) by subtracting known areas of these habitats, ie landfills, quarries, rocky shores and artificial rocky/hard shorelines (delineated by existing topographic data) from the initial set of areas with the same spectral signature. Further manual discrimination was also performed by identifying the highly modified areas of reclaimed land and construction sites (which both belong in the "Other" category) from the bare rock or soil habitat. |
While the habitat was considered as having a high level of mapping accuracy, spot-checking was required at some identified areas. There was little existing information on individual bare rock and soil sites with which to adjust this indicative ecological ranking. |
A total of 14 locations of bare rock or soil habitat were surveyed and their habitat type and boundary confirmed. |
There are about 1,441 hectares of bare rock or soil not covered under the present study. As the majority of the sub-sites surveyed have been mis-classified, further efforts may be devoted to surveying other bare rock or soil areas in an attempt to pick up mis-identified areas. |
| Grassland |
Once the initial, unsupervised classification was complete, aerial photographs were examined, with the assistance of HKU specialists, to select areas of known grassland. In addition, preliminary site visits added additional known grassland areas to the knowledge base. |
While the habitat was mapped with assistance from terrestrial ecology specialists and additional site information from preliminary site visits, uncertainties remained in some of the identified areas. As few studies have investigated the ecological value of grassland, there was little basis in the existing information on which to adjust the indicative ecological ranking for individual sites. |
A total of 10 patches (sub-sites) of grassland were surveyed and the habitat type and boundary were verified. |
There are about 26,095 hectares of grassland left unsurveyed by the present study. As a number of the sub-sites surveyed were mis-classified, further efforts may be required to survey other grassland areas in an attempt to pick up mis-identified areas. |
| Modified Watercourse |
The habitat was mapped by using 1:1K digital topographic data from the Lands Department. EPD (1998) River Water Quality in Hong Kong was used to justify the necessity to further adjust natural watercourses into this category based on their pollution status. |
Despite the detailed Lands Department data, some of the junction areas for modified and natural watercourses could not be clearly identified and thus required field surveys to check the classification of natural and modified watercourse boundaries. Few data on the ecological status of modified watercourses were available apart from the sources from those of EPD? River Monitoring data (1998). |
A total 18 locations (sub-sites) of modified watercourse were surveyed and their habitat type and ecological value verified. |
Apart from those watercourses surveyed under the present study, there are still 95 ha of watercourses yet to be confirmed in terms of their habitat type and boundary. |
| Artificial Rocky/Hard Shoreline |
The habitat was mapped primarily based on manual identification using topographic data and aerial photos. This habitat type was distinguished manually from natural shores based on its relatively straight coastline nature. |
Ecological surveys undertaken on artificial shores were few. Information on those artificial shorelines which may have the potential to be upgraded in terms of their ecological rating was considered lacking. |
Two artificial shores were surveyed under the present study and had their habitat type verified and ecological status justified. |
There are still some artificial shores not surveyed that may have the potential to be upgraded in their ecological ranking. As artificial shores which were engineered as natural boulder or rock habitats and/or have been in place for a long period (eg 10 years) may have the potential to possess higher than low ecological value (as is the case with the High Island Reservoir Dam seawall (Site 156)), further efforts may be devoted to justify ecological value of habitats with this type of structure (ie sloping seawall). |
| Golf Course/Urban Park |
Golf courses and urban parks were mapped by using the 1:20K topographic data from the Lands Department. Little existing information is available to justify the ecological value of golf courses or urban parks. |
As both the location and boundary of golf courses and urban parks were clearly identified based on existing information, their verification was considered not necessary. However, there was little existing information on the ecological status of the habitat and field surveys was required to justify their ecological value. This was particularly necessary for those habitats which contain truly natural and almost intact medium/high value habitats. |
Surveys were conducted on two sites (ie Ocean Park and Chung Hom Kok Park) which are known to contain natural habitats. Ecological status of these two habitats were justified based on the information collected during site visits. |
Sheung Shui Golf Course is one of the additional sites proposed by AFCD (see Table C1 of TR2) for field surveying. As no surveys have been conducted on this site and other identified golf courses, further efforts may be devoted to justify the ecological status of these habitats. |
| Quarry |
The habitat was mapped through manual identification in accordance with the 1:20K topographic data and existing information (Mouchel 1998). Little information is available to justify their ecological rating. |
As the habitat was regarded as having a high level of mapping confidence, only very limited survey effort to spot-check the mapped areas was required. While quarries may have the potential to become areas of conservation interest due to the presence of certain ecologically important species (eg birds of prey), little information was available to substantiate this. |
Three quarries, located at Anderson Road Mount Butler and Shek O, were surveyed and had their boundary verified and ecological value confirmed. |
Lamma quarry is one of the additional sites proposed by EPD (see Table C1 of TR2) for field surveying. As no surveys have been conducted on this site, further efforts may be devoted to justify the ecological status of this quarry. |
|