Assessing Forest Canopy Closure in a Geospatial Medium to Address Management Concerns for Tropical Islands—Southeast Asia

Environ Monit Assess DOI 10.1007/s10661-008-0717-4

Assessing forest canopy closure in a geospatial medium to address management concerns for tropical islands—Southeast Asia

P. Rama Chandra Prasad · Nidhi Nagabhatla · C. S. Reddy · Stutee Gupta · K. S. Rajan · S. H. Raza · C. B. S. Dutt

Received: 9 July 2008 / Accepted: 23 December 2008 © Springer Science + Business Media B.V. 2009

Abstract The present study outlines an approach interpretation for the complete archipelago. In the to classify forest density and to estimate canopy second step, we identified two island groups from closure of the forest of the Andaman and the Andaman to investigate and compare the for- Nicobar archipelago. The vector layers generated est strata density. The third and final step involved for the study area using satellite data was vali- more of a localised phytosociological module that dated with the field knowledge of the surveyed focused on the North Andaman Islands. The re- ground control points. The methodology adopted sults based on the analysis of the high-resolution in this present analysis is three-tiered. First, the satellite data show that more than 75% of the density stratification into five zones using visual mangroves are under high- to very high-density canopy class. The framework developed would serve as a significant measure to forest health and evaluate management concerns whilst addressing B · P. R. C. Prasad ( ) K. S. Rajan issues such as gap identification, conservation pri- Laboratory for Spatial Informatics, International Institute of Information Technology, Gachibowli, oritisation and disaster management—principally Hyderabad 500 032, India to the post-tsunami assessment and analysis. e-mail: [email protected] Keywords Tropical · Islands · Forest · N. Nagabhatla · · · World Fish Centre, Batu Muang, Penang, Malaysia Canopy density South Asia Geospatial Management C. S. Reddy Forestry and Ecology Division, National Remote Sensing Agency, Department of Space, Balanagar, Hyderabad, India Introduction

S. Gupta Tropical rainforests are bestowed with a wide Forest Research Institute, Dehradun, India variety of species richness and diversity patterns S. H. Raza (Alwyn and Calaway 1987; Jacobs 1988). Apart Aurora’s Scientiﬁc Technological & Research from these characteristics, they are recognised as Academy, Hyderabad, India dense forests due to the high density of vegetation formed by the clumped distribution of individ- C. B. S. Dutt Indian Space Research Organization, Department uals and also enormously tangled undergrowth of Space, Anthariksh Bhavan, Bangalore, India of different herbs, shrubs, lianas and climbers. Environ Monit Assess

Detailed and accurate maps of these forests— parameters (forest canopy density, habitat diver- condition and structure—are needed for assess- sity, etc.). The satellite remote sensing is best ment of the flora and faunal biodiversity as well as suited for analysis of canopy closures, as elu- for sustainable ecosystem management (Blodgett cidated by Roy et al. (1994) whilst modelling et al. 2000). The conventional way of ground mon- the biophysical spectral response for forest den- itoring for density estimation can be tedious and sity stratification in evergreen forests of South time-consuming, whilst the use of geographical Andaman division and dry deciduous forests over information system (GIS) and remote sensing as a central India. The study was based on the as- platform for estimating the density of these forests sumption that any alteration in forest canopy is may speed up the process and provide for a more reflected in the crown structure. efficient option (Blodgett et al. 2000). Addition- With inspiration from the above statement, ally, the remote sensing data also facilitate spatial this study illustrates the density analysis and the delineation of vegetation density maps through associated attributes for the tropical Andaman various techniques using satellite imagery in con- and Nicobar (AN) archipelago, the islands char- junction with GIS and phytosociological ground acterised by notable ecological complexity, using data (Chauhan 2004; Prasad 2006). In addition, a three-tier approach. The biodiversity of these these data techniques can be combined for spatial islands is noteworthy and the detail assessment prediction and modelling of the vegetation’s bio- poses a challenge. Past extraction practices, over- physical properties (Muukkonen and Heiskanen exploitation of natural resources, and population 2005). pressure has seriously affected the ecosystem sta- Forest canopy density is one of the important bility and forest density. Currently, these forests parameters in the planning and implementation are prone to anthropogenic disturbances due to of forest rehabilitation programs (Rikimaru et al. the high immigration rates and the introduction 2002). It has been suggested that canopy density is of various exotic animals like elephants (Rauf an essential parameter to assess and analyse the 2004). This paper, whilst addressing the need to factors affecting forest growth, its regeneration understand the canopy closure or density of forest and to keep a check on management initiatives communities which are critical for a management in gap area plantations and regeneration status planning process, also describes the use of the spa- (Chauhan 2004). Further, as discussed by Blodgett tial tools viz., a combination of earth observation et al. (2000), the multi-spectral earth observation data along with the GIS tools that can pertinently data with appropriate ground measurements can contribute to the process. Towards this end, the suitably extrapolate across a large geographic re- authors have mapped the forest canopy density as gion, and this has significant advantages for for- part of the first tier for the complete archipelago, est management, especially in areas where forests whilst the second tier takes up an in-depth analysis are in remote locations or are inaccessible. Also, of two islands in the Andaman group—North and the spatial techniques to measure the canopy clo- Baratang Islands (BI) and the final tier relates the sure along with canopy height profiling has been phytodiversity and the density stratification for illustrated from different regions using a wide the North Andaman Islands (NAI). array of spatial and spectral resolutions (Roy et al. It is envisioned that the derived thematic out- 1990, 1996; Rikimaru and Miyatake 1997;Katul puts will provide a synoptic understanding of the and Albertson 1998; Harding et al. 2001; Marshall extent of open and closed forests in the island et al. 2002; Singh et al. 2003). landscape. The detailed analysis of the two islands will depict the scenario of forest density individu- Background ally, for each forest type, thereby providing information on forest stock which can enable selective Advancement in geospatial technologies provides logging operations by the forest department. In a medium to evaluate forest cover in the inaccessi- addition, the other outputs may assist in the iden- ble and remote isles. This can also enable a multi- tification of illegal logging sites and the extent of scalar assessment of the associated ecosystem forest fragmentation. Environ Monit Assess

Study area and objectives into different forest density zones. At the second level, the goal is to understand the variation in The Andaman and Nicobar Islands (ANI) spreads the forest canopy closures in a continuous land over the geographic location of 6◦5 to 14◦45 N mass (NAI) vs. scattered islands (BI). These two lat. and 92◦ to 94◦ E long and vertically stretch tiers of the geospatial assignment encapsulate the over 800 km in the Bay of Bengal with a total statistics to decipher the dissimilarity and overlaps area of 8,249 km2. Nicobar Islands are a group of in the forest communities of the ANI in general 28 islands (total area, 1,841 km2) separated from and the two zones in detail both in terms of com- the Andaman group (6,408 km2) by a 10◦ channel. position and health. In the third tier, an effort ANI are one among the nation’s richest biodiver- has been made for in-depth analysis of the forest sity regions and contains a great assemblage of density patterns vis-a-vis phytodiversity for the endemic plant species coupled with high species NAI. diversity and density (Roy et al. 2005; Prasad et al. 2007a; Nagabhatla and Roy 2007). The second tier of the study focuses on NAI (12◦95 Nand 92◦86 E) and BI (12◦18 Nand92◦32 E) of Materials and method AN archipelago; NAI constitutes a continuous land mass (1,458 km2) surrounded by few small Satellite data islands like Paget, Point, Smith, Landfall, Inter- view and Narcondam, whilst the BI is more of Indian Remote Sensing Satellite (IRS) 1C/1D Lin- scattered (small/medium-sized islands separated ear Imaging Self Scanner-III (LISS III; 23.5-m res- by straits) landscape with 28 islands, among which olution), Panchromatic (PAN; 5.8-m resolution) the chief are Baratang, Evergreen, Colebrook, and merged LISS-PAN (multi-spectral 5.8 m) Spike, Havelock, Peel, Wilson, Henry Lawrence, datasets were used. Wherever IRS-1C LISS- John Lawrence, Outram and Neil. The islands III cloud-free data were not available, Landsat- have been selected with the purpose of studying Thematic Mapper (TM) data were schematically the variance in density and diversity in two varied referred. The datasets used in this study are listed landscapes. in Table 1 with relevant spectral and spatial res- Predominantly five major vegetation types, viz. olution. Further, the reconnaissance field survey evergreen, semi-evergreen, moist deciduous, man- was undertaken to collect ground truth informa- grove and littoral, flourish in these islands. It has tion about the patterns of vegetation in the area— been observed that the forests of these islands traverses along all roads and major drainage chan- have undergone significant changes during the nels, hill tops, creeks and sandy beaches. To com- last few decades due to large-scale in-migration, pliment the analysis and create a compilation of extensive logging, land conversion and forest the existing knowledge base, literature survey was degradation (details are in Anonymous 2003; carried out and also interaction with forest depart- Nagabhatla et al. 2006; Prasad et al. 2009). To ment and local institutions was established. understand and evaluate the status of the forests before the tsunami struck these regions, the current work analyses the canopy closure. Such an Processing analysis along with the relevant data can prove as a significant baseline to further study the impacts Visual interpretation approach was effectively of tsunami in the post-disaster years. utilised in this study, which optimised full poten- The canopy closure analysis in this work has tial of different band combinations and image en- adopted a three-tier approach which will provide hancement in the multi-spectral data. The image insights into the regional, inter-island and local enlargement techniques in the PAN data added characterisations of the ANI archipelago. In the to the feature delineation. Visual interpretation first level, the entire ANI is analysed using the technique was opted in this study to capture the visual classification approach to stratify the region variability in the complex tropical formations. In Environ Monit Assess

Table 1 Tabulation of multi-spectral and multi-sensor image data used for delineating the forest canopy for the Andaman and Nicobar Islands (Anonymous 2003) Union territory Satellite Path Row Date of acquisition Andaman Islands IRS IC–LISS III 114 64 Mar. 2000 Andaman Islands IRS IC–LISS III 114 65 Feb. 2000 and Mar. 2000 Andaman Islands IRS IC–LISS III 114 66 Feb. 2000 Andaman Islands IRS IC–LISS III 115 63 Jan. 2000 Andaman Islands IRS IC–LISS III 115 64 Jan. 2000 and Mar. 2000 Andaman Islands IRS IC–LISS III 115 65 Jan. 2000 Andaman Islands IRS IC–LISS III 115 66 Jan. 1999 Andaman Islands Landsat TM 134 51 Mar. 2001 Andaman Islands Landsat TM 134 52 Feb. 1997 Andaman Islands IRS IC–PAN 114 64 Feb. 1998 Andaman Islands IRS IC–PAN 114 65 Feb. 1998 Andaman Islands IRS IC–PAN 114 66 Feb. 1998 Andaman Islands IRS IC–PAN 115 63 Feb. 1998 Andaman Islands IRS IC–PAN 115 64 Feb. 1998 Andaman Islands IRS IC–PAN 115 65 Feb. 1998 Andaman Islands IRS IC–PAN 115 66 Feb. 1998 Nicobar Islands IRS IC–LISS III 115 67 Feb. 1999 Nicobar Islands IRS IC–LISS III 116 68 Dec. 1999 Nicobar Islands IRS IC–LISS III 116 69 Aug. 1999 Nicobar Islands IRS IC–LISS III 117 69 Mar. 1998 Nicobar Islands Landsat TM 132 55 Mar. 2000 Nicobar Islands Landsat TM 133 54 Feb. 2001 Nicobar Islands IRS IC–PAN 115 67 Feb. 1998 Nicobar Islands IRS IC–PAN 115 68 Feb. 1998 Andaman and Nicobar IRS IC–PAN and Mar. 1998, Sept. 1999, Islands IRS IC–LISS/PAN merged April 2001, Feb. 2002

the case of the automated classification methods, that pan-sharpens the high-resolution PAN im- the intermixing of closely related forest commu- agery with lower resolution multi-spectral data. nities was relatively high, resulting in low classifi- The technique extracts edge information of high- cation accuracy due to limited spectral bandwidth resolution data, thereby adding it pixel by pixel (80–100 nm) available with IRS 1C/1D satellite to low-resolution data (Rokhmatuloh et al. 2003), data (Prasad et al. 2007b). Hence, visual tech- and exhibits minimal distortions in the spectral niques were used coupled with intense ground characteristics of data compared to other image verification. The field knowledge base along with fusion techniques (Chavez et al. 1991). the ancillary data, viz. road, settlement, division At the onset of the process, the LISS III data boundaries, supported the delineation process. were geometrically rectified with reference to Sur- However, we do recognise that the possibilities vey of India toposheet, and the PAN data were of upscaling this approach can be complex. The rectified using LISS III by image-to-image rec- canopy density was primarily mapped using grey tification process. A detailed vegetation (forest tones of PAN data, whilst fused composite image and non-forest) map with various land use/land (LISS-III/PAN merged) shored up the refinement cover features along with predominant forest and verification process. The method applied for types was prepared using LISS III data by on- merging the LISS-III and PAN was the high pass screen visual interpretation technique for ANI filter resolution merge algorithm embedded in (Anonymous 2003), NAI (Prasad et al. 2007b) the ERDAS digital image processing software and BI (Nagabhatla and Roy 2007). The forest Environ Monit Assess

Table 2 Levels of density classiﬁcation in Andaman and Nicobar Islands Level % Canopy cover Category Density class D5 > 80 Very high Very dense, undisturbed intact patches D4 60–80 High Forest extracted once, now well protected and managed D3 40–60 Medium Average forest subjected to disturbance D2 20–40 Low Disturbed and extracted D1 < 20 Very low/Degraded Degraded (natural or anthropogenic processes)

community layer was extracted from the above Results and discussion thematic output and was used as base map to derive density zones for each of the forest types. Vegetation-type-wise density map The density typology described in Table 2 shows the zones based on percentage of crown cover, The ground information, topographic maps and each at an interval of 20%. The primary forest the digital elevation map supported the interpre- types were classified into five density levels by tation of the earth observation data for density visually delineating the PAN data based on spec- stratification. The density zones also describe the tral reflectance and characteristic of canopy cover. status of forest resources in the islands. The clas- The delineated canopy density layer was refined sified output (vegetation map) derived from the using the merged IRS LISS-PAN data. For delin- LISS III data show five dominant types, viz. ever- eating the density zones, the map was enhanced to green, semi evergreen, moist deciduous, littoral 1:30,000, compared to 1:50,000 used in the forest and mangrove forest, along with other classes like cover mapping, as the close cover of the crown degraded forest agriculture, barren, mudflats, etc. caused intermixing and made it difficult to extract (details of the classification can be refereed from the accurate information. Anonymous 2003; Prasad et al. 2007b; Nagabhatla and Roy 2007). The PAN data being monochromatic add information for density zoning taking Field inventory into account the variation in tone (grey...black) andtexture(smooth...speckle)formedfromthe Remote-sensing-based forest-type strata in con- sparse and clumped nature of stems on the real junction with topography and climate provided a ground. spatial framework for ground sampling. Ground control points were distributed based on proportion of the forest classes and were used to collect, Andaman and Nicobar Islands verify and validate the overall approach. For NAI, phytosociological data (60 plots of 0.1 ha) rep- The density stratification reflected that nearly resenting different density classes were collected 37% (1,738 km2) of the forest area in An- and a field inventory was prepared. About 15 plots daman Islands falls under high density (Table 3; (32 × 32 m) were laid out in each density class Fig. 1). Semi-evergreen, moist deciduous and lit- (D1–D4), whilst D5 class was not sampled due toral forests occur mostly as medium density to difficulty in accessing the area. Within each zones. They are subjected to high anthropogenic plot, measurements of girth at breast height were and recreational pressure, and this high human made for all trees having a girth of more than influence is attributed to their peripheral place- 30 cm at breast height along with its height. Data ment in the landscape. Very low-density areas were analysed for stem density, basal area, species have been classified under degraded forests. richness (the number of species recorded in sam- The terrain and accessibility conditions in the pled area), diversity using Shannon diversity index Nicobar Islands have kept the area comparatively (Shannon and Weiner 1963) and important value free from disturbance, and therefore, high den- index (IVI)—a sum of relative density and relative sity (86%, 1,219 km2) is maintained in general. basal area (Blair and Brunett 1976). These islands have dense and intact forests on Environ Monit Assess

Table 3 Area statistics showing the forest canopy density in major vegetation communities of Andaman and Nicobar Islands (derived from Anonymous 2003) Forest classes % Area covered under each forest class Very high High Medium Low Andaman Tropical evergreen 8.96 36.23 29.42 5.39 Tropical semi evergreen 14.56 34.95 38.46 12.03 Tropical moist deciduous 10.69 40.25 37.46 11.61 Mangroves 58.41 37.23 2.80 1.57 Littoral forest 0.27 31.70 46.08 21.95 Nicobar Tropical evergreen – 94.65 4.16 1.18 Mixed evergreen – 38.80 49.77 11.42 Moist deciduous – 8.68 73.87 17.45 Lowland swamp – 89.38 7.94 2.68 Mangroves 100 – – – Littoral forest – 32.13 52.01 15.86 < 20% (very low) density class was mapped as degraded forest (Refer Anonymous 2003) the western coast, as settlements and commu- gradients and easy approachability that allows nication are restricted to the eastern side. The frequent anthropogenic disturbances. An overall salt-tolerant mangrove ecosystem found mainly in observation of the density map obtained depicts tropical and subtropical intertidal regions which the domination of D3 (40–60%) and D4 (60–80%) show very high density reflect the low or almost density zones in the NAI forests. no disturbance to them. Baratang Islands The density pattern of the for- NAI versus BI est formations in BI show that the vegetation communities here have been under serious an- Andaman and Nicobar Islands The canopy den- thropogenic influence. Combining the major veg- sity map shown in Fig. 2 exhibits major portion of etation types with the density classification shows NAI (506 km2) under dense class (with 60–80% that most of the region has canopy cover of 60– canopy cover), thus suggesting high intact con- 80% (Table 4). The factors viz. terrain, accessi- dition of the forest. The very low-density zone bility and non-proximity to a disturbance source covers an area of 75.3 km2, indicating low for- have helped to preserve higher densities in remote est fragmentation rate. These areas were also re- areas and uninhabited islands like John Lawrence, garded as degraded forests and mostly occur near Henry Lawrence, Outram and North Passage. It the habitations or settlement areas (like Diglipur, is observed that about 60% of evergreen forests Ramnagar, Radhanagar, Kalipur, Durga nagar, and majority of mangrove forests (77.16%) are etc.) or along the road sides. The accessibility to under high- to very high-density level. The semi- these areas allows for higher human lopping prac- evergreen and deciduous forests are categorised tices (Table 4). The D4 density class (represented as medium to moderately dense. The overall pic- by major portion of the study area) occurs pri- ture of canopy closure in the region reflects that marily in mangrove (79.7%), followed by littoral Havelock and Neil Islands have been exposed to (43.9%) and evergreen (37.1%). Both the man- large-scale extraction and increase in secondary grove and evergreen forest types display high pro- formations. The Ritchie’s group of islands shows portion of dense area with undisturbed conditions more or less homogenous vegetation with less existing at various spatial levels, i.e. evergreen on human interference. high altitude areas and mangroves near inaccessi- During the past years, the forests have been ble coastal areas. The low density of canopy cover exposed to heavy extraction activities resulting in observed in moist deciduous is due to topographic low density in Baratang and Havelock Islands. Of Environ Monit Assess

Fig. 1 Density stratiﬁcation for major vegetation communities in Andaman and Nicobar Islands, also highlighting the areas for in-depth analysis in Tier2 (Anonymous 2003)

the total, about 32% area is inhabited with open National Park (John Lawrence, Henry Lawrence to medium dense forest. It is observed that low and Outram) show comparatively high canopy to medium dense forests are concentrated around density due to the protected status of the speciﬁed habitations and along the road. The general trend islands. shows that the inhabited islands depict an average The NAI zone falls in the average forest density density between 40% and 60%, whereas the un- between 40% and 80% slotted into two categories inhabited islands have comparatively larger area of D3 and D4, whilst the BI falls more between under high-density forest cover (see Fig. 2). The 40% and 60% in the main Baratang stretch. The islands covered under the Rani Jhansi Marine other islands in the group, of which only few are Environ Monit Assess

Landfall Island

Outram Island

Henry Lawrence Island

Paget Island

Baratang Island

Peel Island

John Lawrence Island

Havelock Island

Sound Island Neil Island

Interview Island

North Andaman Islands Baratang Islands

Fig. 2 Forest canopy closure analysis for North Andaman and Baratang Islands (Anonymous 2003) inhabited, sum up the average density cover of Density stratiﬁcation vis-a-vis phytosociological D4 zone. The reason being that most of these analysis: the case of NAI are designated under reserve forest, tribal reserve and/or wildlife sanctuary, although the inacces- Satellite-derived density classes were sampled for sibility and its isolation adds to the cause. The phytosociological data and an attempt was made made-up pattern and the placement of the islands to analyse and correlate the phytodiversity with in the archipelago also explains the variation, as the density classes as illustrated in Table 5. NAI is a continuous land mass connected to the other islands of the Andaman group by Andaman Trunk Road (ATR), thus facilitating people to Species richness settle around in contrast to the less chances of settlement penetration in the internal dense for- Species richness was high in D1, decreased in D2 est regions. Therefore, the human disturbances and again increased in D3 and D4 classes. The mainly occur along the wayside. Whilst in the case high species richness observed in D1 is because it of BI, the small scattered islands (such as Neil, is an area with highly disturbed community and Havelock, etc.) provides a probability to settle large canopy gaps. This is similar to earlier ob- along the accessible coastlines in addition to the servations that in natural conditions, disturbances settlements along the ATR in the main stretch of either caused naturally or anthropogenically play Baratang. an ideal role in enhancing the species richness Environ Monit Assess

(Köhler and Huth 2007). Hubbell et al. (1999)

20%) studies on plant species richness at Barro < ( a Colorado Island, Panama show high species richness in the quadrants containing gaps than ) Area (%)

2 non-gap quadrants. Further anthropogenic disturbance coupled with fragmentation had a stronger

distributed in all forest and more immediate effect in reducing native 2 species richness and increasing exotic species richness (Karen et al. 2002). Light acts as one of the potential determinants of species response and distribution in an ecosystem (Anderson and Leopold 2002). Disturbances in forest modiﬁes

) Area (%) Area (km the niche by creating gaps that favour the growth 2 of stranger species surviving under open canopy gaps, i.e. heliophytic species that need sunlight (Denslow 1995; Denslow et al. 1998) or shade- intolerant species (Runkle 1982; Barik et al. 1992) in addition to the native species, thus increasing species richness. The zones D3 and D4 that are regarded as high-density communities with varied heterogeneous forest types (like semi-evergreen

) Area (%) Area (km coexisting with evergreen) reﬂect high species 2 richness. On the other hand, D2, which is a partially disturbed community, shows transitional species richness.

Species diversity ) Area (%) Area (km 2 Diversity was not signiﬁcantly correlated with forest density (Davidar et al. 2005), and there ex- ists a weak relationship between species diversity and density classes. Generally, species diversity is linked with the proportion of species as well as the number of individuals represented by each 80%) High density (60–80%) Medium density (40–60%) Low density (20–40%) Very low density

> species. The high species diversity observed in D1 class could be attributed to the existence of open canopy gaps. The regeneration niche the- ) Area (%) Area (km 2 ory (Grubb 1977) and gap partitioning hypothesis (Ricklefs 1979) states that forest canopy gaps help the formation of microsite heterogeneity for the understorey community and provide gradient physical conditions for species partition, resulting in increased species diversity (Anderson and Leopold 2002). The other factor that is respon- Aerial statistics of the density classiﬁcation in North Andaman and Baratang Islands sible for the species diversity variation could be the association of different forest types under each EvergreenSemi-evergreenMoist deciduous 28.9Littoral 39.3 15.5MangrovesTotal density 6.41 12.68Evergreen 6.88 2.3 94.5 8.4Semi-evergreenMoist deciduous 43.64 124.3Littoral 115.0 19.91Mangroves 0.82 57.6 17.38 1.00Total density 1.06 27.56 37.11 2.13 43.28 107.65 0.13 25.53 187.9 0.00 21.2 228.4 506.0 104.3 79.05 104.6 33.64 83.25 79.72 17.82 0 43.90 50.65 33.66 65.89 56.64 46.35 40.15 69.56 13.8 13.4 222.34 40.06 44.5 2.07 39.7 45.02 43.52 464.6 26.99 29.2 27.88 5.84 28.86 71.32 37.76 12.81 9.88 15.85 25.71 12.93 11.6 5.2 11.6 24.8 1.11 129.03 9.35 21.58 14.23 18.7 14.23 10.84 4.91 130.2 23.51 3.74 5.50 4.19 3.47 1.94 8.30 4.6 0.00 20.1 – – 0 – 2.26 39.75 0.00 8.53 – 75.3 5.17 – 2.10 For Baratang Island, the very low density class was characterised as degraded forest, and the total area of this class amounted to 2.10 km Density classesForest typesNorth Very Andaman high Islands density ( Area (km Baratang Islands a classes Table 4 density class. Environ Monit Assess

Table 5 Density-wise Density class <20% 20–40% 40–60% 60–80% phytodiversity analysis for North Andaman Parameters D1 D2 D3 D4 Islands Species richness 125 100 112 117 Species diversity 6.1 5.6 5.8 5.9 Basal area/ha 57.79 50.76 45.03 51.08 Stem density/ha 640 689 726 806

Basal area and stem density Conclusion

Generally, the number of stems per unit area and Thematic information on forest vegetation and basal area are considered important parameters to density derived from the advancements in remote measure density (Roy et al. 1996). The basal area sensing systems contribute pertinently towards decreased from D1 to D3 classes and increased in enriching the relevant scientific data in manage- D4. Higher basal area in proportion to low stem ment planning activities and related framework. density, as seen in D1, indicates the presence of Band 4-3-2 in LISS-III and band combinations stems with larger girth occupying more ground 5-6-3; 5-6-4; 4-3-2 of Landsat TM facilitated the volume. The class D1 can be treated as open on-screen delineation of complex forest strata. canopy system where adequate inflow of energy The spatial details from the monochromatic PAN nutrients (i.e. rainfall and sunlight) takes place images added more spatial details in the out- and form suitable sites for dispersal as well as put maps. This supports the observation that the germination of seeds favouring good growth of multi-spectral range and high resolution are ap- species. This type of situation enhances the species propriately useful for assessing the canopy den- richness, girth and height of stems, making the sity in the complex tropical forest communities. forest congeal to support higher species diver- In addition, the geospatial analysis indicates that sity. It was observed that the stem density show the spectral and spatial characteristics derived by enhancing trend from low (D1) to very high in merging the LISS and PAN images prove efficient highly dense category (D4, D5). Also, the increase for density stratification. Based on the statistics in number of stems is inversely proportional to derived from the remote sensing data, it can be the disturbance, i.e. the lower the disturbance, the concluded that there are serious repercussions higher is the stem density (Bhuyan et al. 2003). of anthropogenic impact on the health of forest which can be attributed to the conversion of one land cover to another due to increased demand for Height of the stem land resources. Such in-depth information would provide valuable input for the prioritisation and As the tree canopy plays an important role in the categorisation of areas for controlled silvicultural density interpretation, it is interesting to identify practices. Besides, the study was a pioneering the species present in each density class forming effort to explore the spatial processes in this canopy top. In D4 class, the canopy top species are forested landscape, including associated commu- evergreen species, and other classes show a com- nities and habitat resources. bination of both evergreen and moist deciduous. In AN archipelago, the mangrove ecosystem Pterocarpus dalbergoides, Dipterocarpus gracilis depict very high-density canopy closure. On the and Bombax insigne formed the important top other hand, heavy extraction has been the major canopy species represented in more than two den- cause of low density in moist deciduous forests. sity classes. Based on the maximum height of the The overlaying of density layer with the dis- species, the canopy top species identified in each turbance level excogitates that the intact areas density class along with the predominant species classified as high-density regions (>60% canopy (based on IVI) forming the main composition are closure) have low disturbance level. As a gen- shown in Table 6. P. dalbergoides represented as eral trend, the low-density areas are closer to a dominant species in all the four density classes. the disturbance source (road, settlements and Environ Monit Assess

agriculture lands), whilst the intact high-density zone is concentrated on hilltops and inaccessible areas. In addition, the analysis for NAI reflects that the assessments of phytodiversity in different density classes do not show notable correlation between phytodiversity and density classes. On the contrary, the analysis suggest that forest-type- wise density analysis is productive in identifying the general pattern of phytodiversity in different density classes, as species richness and diversity changes with the type of forest. In the present analysis, the variation in the phytodiversity parameters among density classes is the grouping of various forest types within each density class; as a result, the relation between phytodiversity and forest canopy density was not very evident. It is suggested that further investigation should focus on forest-type-wise density–phytodiversity analysis whilst documenting the entire forest ir- respective of types. The present analysis supports the study by Dutt et al. (1994) for the island that also promulgated that the closed forests (>40% canopy density) can be considered as ‘conservation zones’ whilst the open forests (<40% canopy density) as ‘forest produce zone’ for careful management. The approach for measuring forest density is valuable whilst managing fragile and unique ecosystems such as mangroves. In this direc- tion, the thematic information extracted and further fitted into ecological models would highlight and define the fundamental parameters governing biodiversity viz. fragmentation, disturbance and biological richness. Moreover, the density stratification provides a baseline study to improvise the existing conservation policies and accord- 20 m)

> ingly prioritising the conservation needs in the isles. Proper management plans and scheduled logging operations that can derive basis from the present analysis can facilitate the stakeholder group, mainly the forest department, to maintain the diversity as well as density of these pristine emerald green islands.

Acknowledgements The study was carried out as part Tabularised description of top canopy and predominant species in varied density zones of North Andaman Islands of project ‘Biodiversity Characterisation in Andaman and sum of relative density and relative basal area of species Nicobar Islands at landscape level using remote sensing = and GIS’ under Jaivigyan-Science and Technology mission Canarium euphyllumDipterocarpus gracilisPterocarpus dalbergoidesArtocarpus chaplashaPisonia excelsaPterocarpus dalbergoidesDipterocarpus gracilisCeltis wightii Bombax insigneArtocarpus chaplasha Amoora Pterocarpus wallichi dalbergoidesAglaia andamanica Terminalia procera Pterocarpus dalbergoides Mitragyna rotundifolia Diospyros oocarpa Pterocarpus dalbergoides Pterygota alata Bombax insigne Aglaia Lagerstroemia Celtis oligophylla wightii hypoleuca Pterocarpus Albizia dalbergoides lebbek Canarium euphyllum Artocarpus chaplasha Diospyros oocarpa Planchonia andamanica Pterocarpus Pterocymbium dalbergoides tinctorium Diospyros pilosula Pterocymbium tinctorium Celtis wightii Pterocarpus dalbergoides Pajanelia rheedi Pterocymbium tinctorium Terminalia bialata Artocarpus chaplasha Dipterocarpus costatus Predominant species (IVI) D1Top canopy or emergent species (height IVI D2 D3 D4 Table 6 project. The effort and the support of the Jaivigyan team Environ Monit Assess members are acknowledged for their contribution in field Dutt, C. B. S. Udaya lakshmi, V., & Sadhasivaiah, A. S. work and analysis. The usual declaimer applies. (1994). Role of remote sensing in forest management. GIS Development. Retrieved 12 February 2008 from http://www.gisdevelopment.net/aars/acrs/1994/ ts5/ts5002.asp. References Grubb, P. J. (1977). The maintenance of species- richness in plant communities: The importance of the regeneration niche. Biological Reviews of Alwyn, H. G., & Calaway, D. (1987). Contribution of the Cambridge Philosophical Society, 52, 107–145. non trees to species richness of a tropical rain forest. doi:10.1111/j.1469-185X.1977.tb01347.x. Biotropica, 19(2), 149–156. doi:10.2307/2388737. Harding, D., Lefsky, M. A., Parker, G. G., & Blair, B. Anderson, K. L., & Leopold, D. J. (2002). The role of (2001). Laser altimetry canopy height profiles: Meth- canopy gaps in maintaining vascular plant diversity ods and validation for closed-canopy, broadleaved at a forested wetland in New York State. The Jour- forests. Remote Sensing of Environment, 76, 283–297. nal of the Torrey Botanical Society, 129, 238–250. doi:10.1016/S0034-4257(00)00210-8. doi:10.2307/3088774. Hubbell, S. P., Foster, R. B., O’Brien, S. T., Harms, K. Anonymous (2003). Biodiversity characterization at land- E., Condit, R., Wechsler, B., et al. (1999). Light-gap scape level in Andaman and Nicobar Islands using disturbances, recruitment limitation, and tree diver- remote sensing and geographic information system. sity in a Neotropical forest. Science, 283, 554–557. Dehra Dun: Indian Institute of Remote Sensing. doi:10.1126/science.283.5401.554. Barik, S. K., Pandey, H. N., Tripathi, R. S., & Rao, P. Jacobs, M. (1988). The tropical rainforests: A first en- (1992). Micro environmental variability and species counter. Berlin: Springer, 295 pp. diversity in tree fall gaps in a sub-tropical broadleaved Karen, A. R., Barry, J. F., & Fox, M. D. (2002). Changes forest. Vegetatio, 103, 31–40. to plant species richness in forest fragments: Fragment Bhuyan, P., Khan, M. L., & Tripathi, R. S. (2003). Tree age, disturbance and fire history may be as important diversity and population structure in undisturbed and as area. Journal of Biogeography, 29(5–6), 749–765. human-impacted stands of tropical wet evergreen doi:10.1046/j.1365-2699.2002.00722.x. forest in Arunachal Pradesh, Eastern Himalayas, Katul, G. G., & Albertson, J. D. (1998). An investiga- India. Biodiversity and Conservation, 12(8), 1753– tion of higher order closure models for a forested 1773. doi:10.1023/A:1023619017786. canopy. Boundary-Layer Meteorology, 89, 47–74. Blair, R. M., & Brunett, L. E. (1976). Phytosociolog- doi:10.1023/A:1001509106381. ical changes after timber harvest in the southern Köhler P., & Huth, A. (2007). Impacts of recruitment in a southern pine ecosystem. Ecology, 57, 18–32. limitation and canopy disturbance on tropical tree doi:10.2307/1936395. species richness. Ecological Modelling, 203(3–4), 511– Blodgett, C., Jakubauskas, M., Price, K., & Martinko, E. 517. doi:10.1016/j.ecolmodel.2006.11.023. (2000). Remote sensing-based geostatistical modeling Marshall, B. J., Wood, C. J., Gardiner, B. A., & Belcher, of forest canopy structure. ASPRS 2000 Annual Con- R. E. (2002). Conditional sampling of forest canopy ference, Washington, DC, May 22–26. gusts. Boundary-Layer Meteorology, 102, 225–251. Chauhan, N. (2004). Mapping, monitoring and modeling of doi:10.1023/A:1013181714844. landscape for biodiversity characterization in Baratang Muukkonen, P., & Heiskanen, J. (2005). Estimating Forest Division, Andaman and Nicobar Islands. PhD biomass for boreal forests using ASTER satellite thesis, University of Pune and Indian Institute of Re- data combined with stand wise forest inventory mote Sensing (Department of Space, Government of data. Remote Sensing of Environment, 99, 434–447. India), India. doi:10.1016/j.rse.2005.09.011. Chavez, P. S., Sides, S. C., Jr, & Anderson, J. A. (1991). Nagabhatla, N., & Roy, P. S. (2007). Measuring land- Comparison of three different methods to merge scape parameters: Fragmentation, disturbance and bi- multi-resolution and multi-sectoral data: Landsat TM ological richness in Baratang Islands (Andaman) for and SPOT Panchromatic. Photogrammetric Engineer- estimating landscape structure, human and environ- ing and Remote Sensing, 57(3), 295–303. ment interlinkages. International Journal of Ecological Davidar, P., Jean Philippe, P., & Egbert, G. L., Jr. (2005). Development, 6(S07), 22–36. http://www.ceser.res.in/ Changes in rain forest tree diversity, dominance and isder/ijed/cont/ijed-s07-abs.html. rarity across a seasonality gradient in the Western Nagabhatla, N., Roy, P. S., & Jagdale, R. (2006). Ghats, India. Journal of Biogeography, 32(3), 493–501. Evaluating the change (1968–2001) in landscape doi:10.1111/j.1365-2699.2005.01165.x. pattern and analyzing disturbance in Baratang Denslow, J. S. (1995). Disturbance and diversity in tropical Forest Division (Andaman Islands), Southeast Asia, rain forests: The density effect. Ecological Applica- GIS Development. http://www.gisdevelopment.net/ tions, 5, 962–968. doi:10.2307/2269347. application/nrm/overview/over003_1.html. Accessed Denslow, J. S., Ellison, A. M., & Sanford, R. E. (1998). 20th August 2008. Tree fall gap size effects on above- and below-ground Prasad, R. C. (2006). Phytodiversity assessment of tropi- processes in a tropical wet forest. Journal of Ecology, cal rainforest of North Andaman Islands using remote 86, 597–609. doi:10.1046/j.1365-2745.1998.00295.x. sensing and GIS. PhD thesis, Osmania University, Environ Monit Assess

and National Remote Sensing Agency (Department of derived from multisensor image fusion techniques be- Space, Government of India), India. tween JERS-1 SAR and Landsat TM data. Geoscience Prasad, R. C., Reddy, C. S., & Dutt, C. B. S. (2007a). Phy- and Remote Sensing Symposium, IGARSS 03 Proceed- todiversity assessment of tropical rainforest of North ings IEEE International, 6(21–25), 3656–3658. Andaman Islands, India. Research Journal of Forestry, Roy, P. S., Diwakar, P. G., Vohra, T. P. S., & Shan, S. 1(1), 27–39. K. (1990). Forest resource management using Indian Prasad, R. C., Reddy, C. S., Rajasekhar, G., & remote sensing satellite data. Asian-Pacific Remote Dutt, C. B. S. (2007b). Mapping and analyzing Sensing Journal, 3, 11–22. vegetation types of North Andaman Islands, India. Roy, P. S., Diwakar, P. G., Singh, I. J., & Bhan, S. K. (1994). GIS development. Retrieved 15 January 2008 Evaluation of microwave remote sensing data for for- from http://www.gisdevelopment.net/application/nrm/ est stratification and canopy characterisation. Journal overview/over004_1.htm. of Indian Society of Remote Sensing (Photonirvachak), Prasad, R. C., Reddy, C. S., Rajan, K. S., Raza, S. H., 22(1), 31–44. & Dutt, C. B. S. (2009). Assessment of tsunami and Roy, P. S., Miyatake, S., & Rikimaru, A. (1996). anthropogenic impacts on the forest of the North Biophysical spectral response modelling approach Andaman Islands. India International Journal of Remote for forest density stratification. GIS Development. Sensing (in press). doi:10.1080/01431160802460070. Retrieved 25 February 2008 from http://www. Rauf, A. (2004). The effect of introduced herbivores on gisdevelopment.net/aars/acrs/1996/ts5/index.shtml. vegetation in the Andaman Islands. Current Science, Roy, P. S., Padalia, H., Chauhan, N., Porwal, M. C., 86(8), 1103–1112. Biswas, S., & Jagdale, R. (2005). Validation of Ricklefs, R. E. (1979). Environmental heterogeneity and geospatial model for biodiversity characterization at plant species diversity: A hypothesis. American Natu- landscape level—A study in Andaman & Nicobar ralist, 111, 376–381. doi:10.1086/283169. Islands, India. Ecological Modelling, 185(2–4), 349– Rikimaru, A., & Miyatake, S. (1997). Development of 369. doi:10.1016/j.ecolmodel.2005.01.001. forest canopy density mapping and monitoring model Runkle, J. R. (1982). Patterns of disturbance in some old- using indices of vegetation, bare soil and shadow. growth forests of Eastern North America. Ecology, 63, GIS Development Retrieved 25 February 2008 1544–1546. doi:10.2307/1938878. from http://www.gisdevelopment.net/aars/acrs/1997/ Shannon, C. E., & Weiner, W. (1963). The mathematical ts5/index.shtm. theory of communication. Urbana, USA: University of Rikimaru, A., Roy, P. S., & Miyatake, S. (2002). Tropical Illinois Press. forest cover density mapping 2002. Tropical Ecology, Singh, I. J., Das, K. K., & Kushwaha, S. P. S. (2003). Forest 43(1), 39–47. stock assessment using IRS LISS III and PAN merged Rokhmatuloh, T. R., Wikantika, K., Munadi, K., & Aslam, data in Timil Forest Range, Dehradun. Journal Indian M. (2003). Study on the spectral quality preservation society of Remote Sensing, 31(1), 49–56.