Summary and Discussion

Brabyn, 2015 stated that central to all landscape research approaches is the need to have a classification of landscape characters so that there will be a frame of reference for communication. In this respect, landscape characterization can be an effective tool to inform the ongoing discourse on the landscape and can provide an integrating spatial framework to give knowledge of distinctive landscapes (Natural England, 2014). This chapter of the research it showed how the landscape of Teknaf could be characterized and recognized as a distinctive and unique character area. This section discusses the variable, method of the characterization, key characteristics of characterized area and circumstances of landscape character types.

Traditionally, landscape classification has been based on natural factors (e.g., climate, geology, soil, elevation) and cultural factors (e.g., ethnic composition, history, archaeology).

But, the cultural factors are complex, and many aspects are difficult to quantify (Romportl &

Chuman, 2012). Furthermore, it is difficult to collect cultural factor data in poor data region like Bangladesh. Therefore, several classifications based on natural or physical factors. For example, a landscape classification was conducted for Saxony (Germany, 18,338 km2) based on only the physical geography (Bastian, 2000). The natural factors used to classify the top 100 rural landscapes of Japan included the land use, average elevation and relief amplitude (Iwata, Fukamachi, &Morimoto, 2011). The classification rules of landscape character area in Hong Kong (China, 1104 km2) were made mainly based on landform, vegetation, land use and urban settlement (City of Hong Kong, 2001). In the present study, physical factors such as watershed area as a unit, landform types and vegetation categories as variable have been used to develop landscape character area of Teknaf peninsula. 33% of the study area occupied by fifty per cent terrain with less than eight per cent slope and elevation is between 90m-150m and 31% area is irregular plain with moderate relief where fifty to eighty per cent terrain has less than eight per cent slope, and elevation is between 30m-90m. So, more than 60% landscape of Teknaf is more than 30m high from sea level and have a steep slope. Considering the vegetation categories, more than 60% area is covered by mosaic and natural forest area. 35 watershed units were identified where only five units are 1332 ha to 2932 ha and others are less than 1048 ha. Due to the mountainous area, most of the watershed units are small in size.

To analyze the physical factors, several different classification methods have been widely used. For example: Principal component analysis (Abu Jaber et al. 2015)(Klinger, Schwanghart, and Schütt 2011), means clustering (Atik, Işıklı, and Ortaçeşme 2016), k-medias algorithm (Iwata et al., 2011; Zhang & Couloigner, 2005), AP algorithm(Li and Zhang

77

2017), Region growing techniques (Jellema et al. 2009), Geospatial tool (Symons et al. 2013).

M.F. Schmitz 2012, divided the landscape of Madrid (Central Spain) into a mosaic of the natural formations with forest land uses and mixed agricultural land systems by using natural factors and principal component analysis. In the study, the similarity between landscape character area was identified using Principal component analysis within ten variables. Among the ten variables, high land and forest area, plain land and grass area, plain relief and homestead garden area are positively related with each other. This analysis extracted two components and four landscape character areas gathered against the two components.

The relationship between vegetation characteristics (species cover and composition) and elevation of the landscape has been documented by (Sánchez, Izco, and Medrano 1996). Table 3.15 is indicating the relationship between landform and vegetation types. Watershed units of plain grass land area and high land forest area declined more than other two LCA. Distribution of high mountainous area is very irregular among all LCA.

Table 3:17 Average percentage of landform and vegetation type are Landform and

vegetation types (area)

Average area (%) and SD (%) of watershed units Plain grass

land

Plain relief homestead

High land forest

Mixed land mosaic

Avg. SD Avg. SD Avg. SD Avg. SD

Plain land (%) 35 12 13 5 13 11 16 6

Plain relief (%) 33 11 31 12 11 9 20 10

High land (%) 25 14 18 8 49 10 33 6

High mountainous (%) 7 9 38 17 27 16 32 10

Betel leaf (%) 0 0 3 2 0 0 2 2

Grass land (%) 41 4 16 2 10 9 11 5

Mosaic (%) 23 10 25 8 20 9 47 6

Forest (%) 24 6 26 7 57 14 31 7

Homestead (%) 7 4 28 6 4 2 6 4

Social forest (%) 3 3 3 3 9 7 3 4

Table 3.16 showing the several distinguishing points of four LCA that indicate the uniqueness of each area which is important for preparing a plan to manage the landscape.

78

Table 3:18 Key characteristics of four LCA

LCA Key Characteristics

Plain grass land  The average size of the watershed unit is 1089.81 ha

 Paddy, vegetable and salt cultivated agricultural field

 68% area having eighty per cent terrain with less than 8% slope

 Timber dominated species in homestead garden Plain relief

homestead

 The average size of the watershed unit is 341.75 ha

 Almost 69% area’s elevation differences very steep

 Paddy, betel leaf and vegetable cultivated the agricultural field

 Supari and fruit dominated species in homestead garden High land forest  The average size of the watershed unit is 560.42 ha

 50% terrain with less than 8% slope and 90-150m elevation.

 Paddy cultivated the agricultural field

 Supari and fruit dominated species in homestead garden Mixed land mosaic  The average size of the watershed unit is 819.62 ha

 Paddy, betel leaf and vegetable cultivated the agricultural field

 Fruit dominated species in homestead garden

Caspersen (2009) emphasized that the grouping of characters of landscapes would be beneficial data to plan and manage cultural and natural landscapes. The four landscape character area identified six landscape character types. From table 3.17 agricultural field dominated all LCA, but betel leaf field is more suitable in plain relief homestead garden and mixed land mosaic area. Except for high land forest area, all other LCA have the almost similar amount of natural forest. High land forest area is highly suitable social forestry.

Table 3.19 presents, the differences of the occupied area of LCT in four LCA. Although, the Teknaf peninsula is a forested landscape but it has a great opportunity to develop a sustainable agricultural system over the enormous area of agricultural field. Betel leaf related management needed in plain relief homestead garden and mixed land mosaic area. Mosaic area mentioned in the table 3.19 actually indicates the degraded area within the natural forest area.

Therefore, through the replantation on mosaic area could enhance the natural forest protection.

All of the LCAs except high land forest area encourage homestead garden. May be the physical character of the high land forest area is not suitable for expanding homestead garden but very much suitable for social forest.

79

Table 3:19 Landscape character types area (%) in four LCA LCT

(Area)

Plain grass land

Plain relief homestead

High land forest Mixed land mosaic

Agricultural field (%) 67 53 29 44

Betel leaf (%) 1 13 1 9

Natural forest (%) 14 7 32 12

Homestead garden (%) 10 20 4 19

Mosaic Area (%) 6 5 25 14

Social forest (%) 2 2 8 2

From the result of statistical analysis and field survey, it is proved that watershed unit, landform and vegetation data sets and method are suitable to identify distinguishing pattern and characterized area of landscape of the study area.

80