On the inadequacy of environment impact assessments for projects in Bhagwan Mahavir Wildlife Sanctuary and National Park of Goa, India: a peer review

The Environment Impact Assessment (EIA) is a regulatory framework adopted since 1994 in India to evaluate the impact and mitigation measures of projects, however, even after 25 years of adoption, EIAs continue to be of inferior quality with respect to biodiversity documentation and assessment of impacts and their mitigation measures. This questions the credibility of the exercise, as deficient EIAs are habitually used as a basis for project clearances in ecologically sensitive and irreplaceable regions. The authors reiterate this point by analysing impact assessment documents for three projects: the doubling of the National Highway-4A, doubling of the railwayline from Castlerock to Kulem, and laying of a 400-kV transmission line through the Bhagwan Mahavir Wildlife Sanctuary and National Park in the state of Goa. Two of these projects were recently granted ‘Wildlife Clearance’ during a virtual meeting of the Standing Committee of the National Board of Wildlife (NBWL) without a thorough assessment of the project impacts. Assessment reports for the road and railway expansion were found to be deficient on multiple fronts regarding biodiversity assessment and projected impacts, whereas no impact assessment report was available in the public domain for the 400-kV transmission line project. This paper highlights the biodiversity significance of this protected area complex in the Western Ghats, and highlights the lacunae in biodiversity documentation and inadequacy of mitigation measures in assessment documents for all three diversion projects. The EIA process needs to improve substantially if India is to protect its natural resources and adhere to environmental protection policies and regulations nationally and globally.


Plants and Fungi
BMWS & NP comprise more than 700 plant species (Datar & Lakshminarasimhan 2013; See Appendix II). Of these, at least 127 species are endemic, making about 18% of the total flora (Datar & Lakshminarasimhan 2011).
The region is a hotspot for fungal diversity in the Western Ghats. Nearly 1,200 fungi species are known from Goa, of which at least 500 mushroom species have been identified and many are yet to be described (Nandkumar Kamat in litt. 27.xii.2020). A total of 18 lichen species are known from the PA, although the overall diversity is likely to be higher.

Insects and Arachnids
Both PAs together support 219 butterfly species (Appendix V) and 80 species of odonates (Appendix VI), of which 14 species of butterflies and 18 species of odonates are endemic to the Western Ghats. Two odonate species Idionyx gomantakensis (Subramanian et al. 2013) and Cyclogomphus flavoannulatus (Rangnekar et al. 2019) Act, 1972(henceforth WPA 1972. Two endemic butterfly species found here are the Malabar Rose Pachliopta pandiyana and the Southern Birdwing Troides minos. A 2011 report on moth diversity from the northern Western Ghats reports at least 418 moth species out of which 116 species were unidentified, and potentially new to science (Shubhalaxmi et al. 2011). A total of 75 ant species are recorded from the WLS of which seven are endemic (See Appendix IX). Six scorpion species, 16 spider species, and one species each of Whip Scorpion and Whip Spider have been recorded from both the PAs (Bastawade & Borkar 2008). An isolated population of Whip Spider Phrynichus phipsoni and Whip Scorpion Labochirus tauricornis occurs in the proximity of this PA (Borkar et al. 2006;Borkar 2018).

Fish
The Western Ghats supports over 300 fish species of which more than 65% are endemic (Kumar & Devi 2013). New fish species and range extensions are being described from this region as yet, suggesting that fish species assessments and distribution patterns remain incomplete (Molur et al. 2011). A comprehensive study in the Mhadei sub-basin (which includes BMWS & NP) has the presence of 49 fish species, of which 18 species are endemic to the Western Ghats (Atkore 2017; See Appendix IV).

Herpetofauna
The reptilian diversity of the region is represented by 52 species from Crocodylidae (Crocodiles), testudines (freshwater turtles & tortoises), and squamates which includes Sauria (Lizards) and Ophidia (Snakes) (See Appendix VII). Amongst the diversity of reptiles, the Indian Rock Python Python molurus, Indian Monitor Lizard Varanus bengalensis, and King Cobra Ophiophagus hannah are some species in the Schedule I and II of WPA, 1972. Other endemics such as the Malabar Pit Viper Trimeresurus malabaricus and the Large-Scaled Shieldtail Uropeltis macrolepis are also reported from the region.
In the past 15 years, 112 new amphibian species have been discovered from the Western Ghats, indicating high species richness and a need for more systematic studies in the landscape. Among the 218 known species of amphibians, 87.8% (158 species) are endemic to the J TT Western Ghats (Nirmal Kulkarni pers. obs. 01.vii.2020). The two PAs together contain at least 36 amphibian species (See Appendix VIII). Castlerock is the type locality of Nyctibatrachus petraeus (Das & Kunte 2005) and Raorchestes bombayensis (Annandale 1919). Biju et al. (2014a) described 14 new dancing frogs, of which one species Micrixalus uttaraghati is found in the streams that cut across the existing Castlerock-Kulem railway line. Similarly, these streams are home to Indosylvirana caesari and Indirana chiravasi, two new frog species that were described recently (Biju et al. 2014b;Padhye et al. 2014). Seven new amphibian species have been discovered in the past two decades from Goa.

Birds
The first ornithological study in Goa was conducted by Grubh & Ali (1976). During their 16-day survey that included Mollem, the team recorded a total of 97 bird species. Presently, 286 species have been recorded from the BMWS & NP (Rahmani et al. 2016; see Appendix I). The list includes species such as the Critically Endangered Indian Vulture Gyps indicus, Endangered Egyptian Vulture Neophron percnopterus and other globally threatened species such as the Lesser Adjutant Leptoptilos javanicus, Woolly-necked Stork Ciconia episcopus, Nilgiri Wood Pigeon Columba elphinstonii, and Malabar Pied Hornbill Anthracoceros coronatus. A total of 14 bird species recorded from BMWS & NP are endemic to the Western Ghats and 32 of the recorded species are listed in the Schedule I (Part III) of the WPA, 1972. Six bird species are classified as Near Threatened by International Union for Conservation of Nature and Natural Resources (IUCN).

Mammals
BMWS & NP, along with the Kali Tiger Reserve and surrounding reserved and protected forests cover an area of at least 2,000km 2 and form an important Bengal Tiger Panthera tigris habitat (Gubbi et al. 2016). The National Tiger Conservation Authority has recommended bringing together the protected areas of Goa and Karnataka for Tiger conservation and improved management. In a document released by the MoEFCC titled "Connecting Tiger Populations for Longterm Conservation" the forests of Goa are mentioned as one (Sahyadri-Radhanagari-Goa) of 32 major Tiger corridors in India. A breeding population of Tigers has been recorded from the tri-junction of Goa-Karnataka-Maharashtra (Girish Punjabi pers. obs. 19.iii.2019;Jhala et al. 2020). In May 2019, the Forest Department of Goa photographed an individual Tiger using trail cameras in the BMWS & NP, and expect more individuals to be present (The Goan Everyday 2019). On the 5 January 2020, carcasses of four Tigers -a female and her three cubs were found in the neighbouring Mhadei Wildlife Sanctuary (Kerkar 2020). The four tigers were poisoned in retaliation for depredating livestock (Kerkar 2020).
More than 60 mammal species are likely to occur in the PAs, of which 11 species belong to Schedule I of the WPA 1972 (See Appendix III). Gad & Shyama (2009) found that Gaur Bos gaurus was widespread and fed on 32 plant species belonging to 17 plant families in the PA. Sengupta & Radhakrishna (2013) encountered a higher number of Bonnet Macaques Macaca radiata in BMWS & NP as compared to other parts of Goa. Krupa et al. (2017) reported two sympatric otter species, the Asian Small-clawed Otter Aonyx cinereus and Smooth-coated Otter Lutrogale perspicillata in the buffer region of Mhadei WS, which adjoins BMWS & NP.

REVIEW OF EIA FOR THE NH-4A HIGHWAY EXPANSION PROJECT
After compiling lists of taxonomic groups known from the PAs, the authors reviewed the EIA for NH-4A Highway Expansion (henceforth Road Report) for information provided on taxonomic groups, environmental and social impacts of the project. We found inadequacies in most aspects and as such the Road Report was observed to be of poor quality. The relevant issues are discussed here.

Plants
i) Several issues were found with the reported methodology for the baseline survey on plant species in the Road Report. The sampling strategy was not clearly indicated. The Report says that the number of quadrats in each habitat type was proportionate to the land in the habitat type, but no further information is provided whether a randomized or systematic sampling protocol was used. In the absence of a protocol, sampling locations would be biased and not fully representative of the habitat type.
ii) Resultantly, the list on floral species in the Road Report is inadequate when compared to existing species list of the area (Datar & Lakshminarasimhan 2011).
iii) The sampling methodology outlined was likely not followed. The data were finally compiled and based on a reconnaissance trip and secondary literature (Section 5.8.2,Page 79). No analysis of species diversity or dominance were performed, and the findings are only provided in the form of a brief species list (Table 5.15,

J TT
Page 79). This list excludes understory species, herbs, and lianas. No data are presented on tree girths, size classes or age structure, which could help in an assessment of the damage to the forest.
iv) The Road Report is not clear about which agency will plant trees as part of the project and the figures provided are inconsistent. It states that the intention is to plant 20,000 avenue trees next to roadside parking areas, toll gates, bus bays, and truck lay-bys (page 82). The Report later revises this number to 50,000 trees (Page 96), then to ~27,000 trees (based on planting 333 trees per km 2 in wildlife sanctuary and 666 trees per km 2 elsewhere) and finally back to 20,000 trees (Page 137).
v) The Road Report has also not specifically identified plant species that will provide appropriate and compensatory ecosystem services for the tree plantation. The species to be planted on the edge of the highway are Mahua Madhuca longifolia and Khair Acacia catechu, which are not typical of the Western Ghats and Bougainvillea sp., an ornamental that is not native to India, but is to be planted in the median.

Insects and Arachnids
i) The Road Report has no impact assessment of linear intrusions on insect and arachnid diversity, with no details on species richness in the area. No attempt has been made to compile secondary information from published and unpublished sources.
ii) There are many studies from India and the world which have examined the impact of roads on insects. Insects suffer a high mortality while crossing roads or may avoid crossing roads altogether (Muñoz et al. 2015). Studies report that vehicle lights attract many insects, causing mortality during the night (Seshadri & Ganesh 2011). The barrier effect of roads is higher for slow-moving insects (Muñoz et al. 2015), but even flying insects such as butterflies are affected by fragmentation created by roads, as the nature of patch-edge affects their dispersal ability (Ries & Debinski 2001;Dover & Settele 2009). Studies on grasshoppers have demonstrated that males increase their call frequency in response to road noise, which may have population-level consequences (Lampe et al. 2014).
iii) Despite evidence of high levels of diversity and endemism in the BMWS & NP, odonates have not been surveyed. Species of hill streams are more narrowly-distributed and are indicators of water quality (Simaika et al. 2016). A new dragonfly species Idionyx gomantakensis (Subramanian et al. 2013) was reported in 2013 from the PAs, a fact that has been overlooked in the Road Report. This raises doubts whether secondary data has been reviewed while compiling the faunal list for the project area. iv) No details are provided for Arachnids in the Road Report. In so far as the amblipygid, urropygid, and theraphosid spiders of these areas are concerned, given their fidelity to their habitat type and their rather restricted movement, any alteration of the habitat, due to road construction and widening, shall decimate these small and isolated populations beyond recuperation and renewal, even leading to local extinctions (Maelfait & Hendrickx 1998).

Fish
i) No fish species or impacts of road expansion have been described in the Road Report. It states that "since most of the water bodies remain dry during the non-monsoon months, this [sediment] impact will be negligible" (Page 98). This statement is inaccurate, as several perennial streams and pool habitats contain water and act as refuges for various fish species in the non-monsoon months. A study cautions against the effects of sedimentation and run-off on the fish communities due to rampant vehicular traffic in the neighbouring Mhadei Sanctuary (Atkore 2017).
ii) Many other impacts are envisioned which the Road Report has not assessed. Soil erosion due to the removal of riparian vegetation would have shortterm as well as long-term impacts on stream dwelling communities. Riparian vegetation plays an important role in maintaining ambient temperature in the headwater catchment (region from numerous streams originates) enabling persistence of diverse, endemic and habitat specialist fish species such as Balitora sp., Glyptothorax sp., Schistura sp., Bhavania sp., and Garra sp. (Sreekantha et al. 2007).
iii) Increased soil erosion due to the road expansion is likely to multiply the sedimentation load, which may impair water quality greatly due to high turbidity. Sediment deposition is likely to reduce food availability to aquatic communities. Bottom-dwelling fish such as Balitora sp., Glyptothorax sp., and Schistura sp. feed on benthic insects (Daniels 2002), and have a very narrow range of distribution and tolerance level to certain water quality variables. iv) Higher suspended solids and silt deposition can also affect spawning grounds and various life stages of fish. A few highly sensitive fish species such as Deccan Mahseer Tor khudree and Hypselobarbus sp. are known to migrate upstream for feeding and breeding, either once or twice a year. Mahseer, in particular, are known to choose definite and special spawning grounds which J TT usually are rich in dissolved oxygen content, neutral pH, and cool water temperature. Eggs, fry and fingerlings stages of this fish are highly sensitive to the slight alterations in their environment and spawning habitats (Daniels 2002). Soil erosion and high deposition of silt along with stream flow are expected to destroy their habitat, and could reduce their population in Dudhsagar and other adjoining waterbodies. v) Surface dwelling fish such as Devario sp., Barilius sp., and Salmostoma sp. feed largely on insects falling from the canopy (Johnson & Arunachalam 2010). Higher turbidity due to sediment load would reduce their ability to forage and may restrict these fishes to downstream habitats, affecting their survival.

Herpetofauna
i) Details on herpetofaunal diversity in the PAs is not mentioned in the Road Report. Section 5.8 (Page 74) mentions that Goa has a high snake population. While this may be a general statement, it is not backed by any references.
ii) Further, data from exsiting literature points to an increase in the number of snake and amphibian road-kills with existence of roads (Garriga et al. 2012;Santhoshkumar et al. 2017). There is, however, no mention of the impact of road expansion on herpetofaunal diversity of the PA in the Road Report.

Birds
i) Although the Road Report mentions that a field survey has been carried out (Section 5.8.2 (v), Page 78), there is no bird checklist provided, except for one mention of the Indian Robin Copsychus fulicatus along with other fauna (Table 5.16, Page 82). Bird species richness and abundance were not quantified in the project area that may be affected due to the project construction. This is a serious shortcoming given that 286 bird species have been recorded in the BMWS & NP (Rahmani et al. 2016, See Appendix I).
ii) A section (Page 74) of the Road Report matches the Wikipedia page "Flora and Fauna of Goa" (Wikipedia contributors 2020), which mentions that the state bird of Goa is "the Ruby-throated Yellow Bulbul, which is a variation of Black-crested Bulbul". This is inaccurate, as the state bird of Goa is the Flame-throated Bulbul Pycnonotus gularis, which recent studies have elevated to a full species (Rasmussen & Anderton 2012).
iii) There is further confusion about the state bird of Goa; section 5.8.1 of Page 78 of the Road Report refers to the Yellow-throated Bulbul Pycnonotus xantholaemus as the state bird. The Yellow-throated Bulbul is endemic to peninsular India and has no known distribution in Goa. The faunal statistics presented in section 5.8.1 have been taken from Kumar & Somashekar (2008) with no attribution to the original source. The absence of any data on birds, either quantitative or qualitative, from an area that has been classified as an Important Bird and Biodiversity Area (IBBA), undermines the purpose of the EIA.

Mammals
i) To assess faunal diversity, field surveys and a local consultation were conducted in the Road Report, however, it does not contain any methodological specifications or sampling strategy. Sampling methods for different taxa are also not clearly differentiated. The species list is limited with only 16 species recorded. This is a clear underestimate as more than 60 mammal species are known to inhabit this region (See Appendix III).
ii) The presence of the Bengal Tiger in the area is also not mentioned. The report states that no endangered species are found in the area which is clearly misleading considering three Endangered mammal species occur, including the Tiger, Dhole, and Indian Pangolin. One of the species mentioned in the Road Report, the Red Giant Flying Squirrel Petaurista petaurista is not found in the Western Ghats. Common species such as the Bonnet Macaque and Chital Axis axis, are also not reported.
iii) The Road Report states that the road expansion will not affect faunal species, and instead claims that species "may increase in number because of the road structures as the project will not obstruct their movement rather can create new habitats for them" (Section 5.8. 1,v,Page 82). This statement is misleading as wide roads are known to create an obstruction to movement for a wide variety of species, including mammals (Bennett 2017). Roads also create forest edges that can harmfully affect native vegetation and rare wildlife due to edge effects, which extend far beyond the area of the road (Gubbi et al. 2012;Poor et al. 2019). Small mammal communities near roads have also been found to differ from those away from roads (Goosem 2002). iv) Section 5.8.1 (Page 78) of the Road Report mentions the Leopard and Black Panther as two separate species, however, these are colour morphs of the same species Panthera pardus. The Gaur Bos gaurus, which is the State Animal of Goa, does not find mention in the checklist. Section 5.8.1, (v) also states that none of the faunal species found here are "endangered or extinct". This is unsound as endangered species such as the Bengal Tiger, Dhole, and Indian Pangolin are found in the region, while extinct species are found nowhere in the wild.

Land-use
i) A land-cover map for this project was acquired as a secondary data source, without clarity on how it was prepared. The map presented is for the entire state of Goa ( Figure 5.23, Page 75), and not specific to the project site. The impacts on the land-use and land-cover specific to the project area have not been assessed in the EIA. The land-use table (Table 5.14, Page 73) has an error in summation of all land-use types. Further, the land-cover classes in the table do not match the ones in the map. These errors create confusion about which land-cover types will be affected by the project.

Water
i) The Road Report mentions that there are declining water level pockets in South Goa, indicating the need to strictly regulate groundwater extraction in these pockets, however, Section 5. 1.4. (Page 40) of the Report has insufficient information on the river basins in the region. Only water depths are provided, without any data on the coverage area, volumes, or a reasonable level of water extraction that is possible from rivers during road expansion.
ii) Section 5.5.1. (Page 51) states that chloride concentrations are "well within the desirable and permissible limits". This statement is misleading. Samples GW-02 and -06 both had detected values above the desirable limit range and are at risk of exceeding the Bureau of Indian Standards' drinking water standards. Thus, there is insufficient evidence to support the statement that there is 'good' scope for groundwater exploitation in all the five affected taluks in the South Goa District.
iii) There is inadequate information on the water assessment sampling procedure in the region. Section 5.5.1 (Page 51) suggests that single samples were collected from five separate surface water sites and eight separate groundwater sites, during one sampling visit. No indication of the season or sampling date is provided, nor of repeated sampling to ensure accuracy or reliability. The statement that "total hardness observed to be constant in all samples" is flawed, as notable variation was observed between hardness in the different sample locations of the Road Report.

Air
i) Air quality would be negatively affected after the road expansion, but there is scarce attention paid to any robust evaluation in the Road Report. The statement that with the "proposed four-laning project, traffic may further come down and ease the vehicles movement and traffic congestion, which may lead to reduce the pollution levels" lacks substantial evidence and cannot be a justification for road expansion within a PA and ecologically-sensitive area.
ii) Table 6.5 (Page 102) proposes that greenhouse gases and other pollutant emissions may be significantly reduced based on the assumption of a small increase in traffic burden along with the avoidance of stopping, idling and congestion, however, traffic projections in the report show that total traffic is projected to only increase over the years, at all the three points where present traffic was surveyed (Table 2.15, Page 20). It is doubtful that vehicular emissions will be reduced with increased number of lanes, when scientific literature indicates that road widening leads to increased emissions which negatively affect air quality (Roberts et al. 2010;Font et al. 2014) 3.10 Soil i) According to Table 5.13. (Page 72) of the Road Report, among the trace metals likely to contaminate soils due to large-scale construction and traffic pollution, only Lead (Pb) and Iron (Fe) are noted, however, this is this is insufficient, as several heavy metals such as Cadmium (Cd), Copper (Cu), Zinc (Zn), and Manganese (Mn) originate from material abrasion, fuel combustion and road dust (Chen et al. 2010;Abdel-Latif & Saleh 2012;Świetlik et al. 2013). Heavy metals have been associated with high levels of genotoxicity and mutagenicity in soils contaminated with heavy metals (Husejnovic et al. 2018) and their concentrations should be monitored and potentially reduced in PAs, particularly in view of the risk of trophic transfer, migration, and bioaccumulation (Zhang et al. 2018;Chouvelon et al. 2019).

Social Impacts
i) Datar & Lakshminarasimhan (2011) documented around 90 floral species to be important for local consumption and livelihood. While the Road Report lists flora of the affected area and people's reliance on non-timber forest produce (NTFP), it does not mention the potential impacts on the floral community that can hamper NTFP-based livelihoods of the local community around BMWS & NP.
ii) The Road Report mentions that apart from forest land, almost 70.42ha of non-forest land would be acquired affecting 377 civilian and governmental structures (Table 5.17, Page 85). It is not clear what the extent of damage to these structures would be. Further, the assessment does not delve deeper into the livelihood impacts and possible mitigation plans for families J TT affected by the project. It mentions that a separate land acquisition plan would be devised for these aspects and has no concrete mitigation plans for social impacts.

REVIEW OF ASSESSMENT STUDY FOR THE RAILWAY EXPANSION PROJECT
While the assessment study (hereafter Railway Report) for railway expansion was informative and detailed, it suffered from several shortcomings as well. The authors reviewed it for information on the same parameters -assessment of taxonomic groups, environmental, and social impacts. It is noteworthy that the railway expansion will affect not only the BMWS & NP but also the neighbouring Kali Tiger Reserve in Karnataka State. Therefore, a project which will fragment the only intact tiger and elephant population in the north-central Western Ghats will have severe ramifications for wildlife and biodiversity. The Railway Report, however, does not stress on the ecological impacts of railway expansion and instead presents a neutral portrait of the project impacts by emphasizing uncertain mitigation measures.

Plants
i) In the section on vegetation characteristics, it is mentioned that 255 species of flowering plants were recorded (Page 64), but Appendix 2.1.a. of the Railway Report lists 224 woody trees. The IUCN Red List status is not provided, and a few common endemic species that occur in the region are not mentioned in the tree species list.
ii) The floristic survey results (Page 83) only records seedlings of woody trees but not herbs and orchids, some of which are rare with restricted distribution in the Western Ghats (Joshi & Janarthanam 2004).
iii) Plant species are misspelled or outright erroneous in the Appendix which makes it difficult to identify the plants that will be impacted. For example, Euonymus undulatus is misspelled (correct name: Euonymus angulatus), while Lapisanthes microphylla is an invalid scientific name as per our knowledge. iv) Appendix 2.1.a of the Railway Report mentions 13 plant species (including vulnerable and endemic species) which are yet to be recorded from Goa. Three of those species may not occur in the BMWS & NP and need further scrutiny as to the validity of their inclusion, however, even if they do occur, it only reveals the importance of the region for plant diversity, and therefore the region should not be diverted for the railway expansion.

Insects
i) The Railway Report follows standardized protocols to document butterfly diversity of the region but covers a very small area which might not represent all the habitats affected by the project, a fact acknowledged in the study (Page 87).
ii) The survey was carried out from April (2013) to May (2014), however, there is no mention of the duration of data collection, including details on whether surveys were undertaken every month or a few days every season. This would have a bearing on the findings.
iii) There is no mention of whether sampling effort was replicated. This precludes an understanding of how many times a transect was sampled, and whether the same transects were sampled repeatedly in subsequent seasons. Quantitative analysis of data collected with inadequate sampling protocols may lead to incorrect estimates of insect diversity.
iv) The Railway Report mentions that the Family Lycaenidae and Hesperiidae were represented by 33 and 18 species, respectively. The number of species, however, might be under-represented given the difficulty in visual identification of species belonging to these Families. No effort was made to account for detection issues in the Railway Report.
v) The Railway Report also does not provide an assessment on moth diversity. Moths are ecologically important and even more diverse than butterflies and dragonflies. At least 418 species of moths of which 116 species are unidentified, were reported from the north Western Ghats (Shubhalaxmi et al. 2011). Given that the study site is a PA in the Western Ghats, it is likely to have high moth diversity.
vi) There are discrepancies in the listing of species in the Railway Report. For example, butterfly species such as Neptis columella, Doleschallia bisaltide, Actolepis puspa, and Castalius rosimon which are Schedule I species are left out of the scheduled species list and the text, with only a passing mention in the Appendix of the Railway Report (Appendix 2, Page 89).

Fish
i) The Railway Report records the presence of 23 fish species, however, a comprehensive study in the Mhadei sub-basin (which includes BMWS & NP) reported 49 fish species with 18 endemics from the Western Ghats (Atkore 2017; see Appendix IV).
ii) The Railway Report does not assess potential impacts of the project on fish community structure, even though studies have found that alteration of stream environment (changes in water quality and flow J TT alteration) by anthropogenic pressures have negative influences on fish guild composition (Atkore 2017; Atkore et al. 2020).

Herpetofauna
i) The Railway Report has a fairly comprehensive assessment of amphibians and reptiles. It reports key details about the diversity of herpetofauna, including endemics, however, it only mentions the impact of the railway-line in causing mortality of reptiles (Page 140), and remains inconclusive of impacts on amphibians (Page 135).
ii) The survey on amphibians clearly finds that 13% of endemic Western Ghats species (14 species out of 24) were found in the project area. This number is likely higher and points to the sensitivity of the region for anurans (See Appendix VIII of this paper).
iii) For reptiles, the Railway Report finds 27 species, which is an underestimate (See Appendix VII of this paper). The report does not have an exhaustive assessment of impacts due to the railway expansion on herpetofauna, reasoning that the study was carried out "during the inactive period of reptiles (winter) where the intensity of the impact could not be assessed properly due to their high seasonal activity, secretiveness and less conspicuousness" (Page 140).

Birds
i) The Railway Report mentions that a two-day survey for birds was carried out in September 2014 and May 2015. It is not clear why a short survey effort was employed to compile the checklist. The survey enumerates only 35 species, of which nine were endemic species. This is an underestimate, compared to the 286 bird species recorded in the BMWS & NP in a comprehensive checklist (Rahmani et al. 2016;eBird 2017).
ii) Data is collected only for cavity-nesting birds. This omits species that do not nest in cavities, but are dependent on trees and vegetation for nesting and feeding. The reason for surveying only cavity-nesting birds is not provided. Further, migratory birds are underrepresented in the survey, given that the survey was not carried out during the migratory season between October-March.
iii) The Railway Report mentions, "The loss of tree specially >10 and >60cm dbh would impact the nesting of birds in the proposed project area" (Page 145). Again, this focuses only on cavity-nesting birds, and undermines the importance of shrubs and undergrowth for passerines and understorey insectivores, which will also be impacted. Such impacts of the loss and fragmentation of the forest cannot be mitigated or compensated for, with respect to ground-nesting and understorey insectivorous birds (Lampila et al. 2005).
iv) The project area description (Page 19) mentions the state bird of Goa as the Ruby-throated Yellow Bulbul Pycnonotus dispar. This is an error. The state bird of Goa is the Flame-throated Bulbul Pycnonotus gularis, while P. dispar is a bird found in the forests of Java and Sumatra.

Mammals
i) The Railway Report suffers from multiple lacunae such as inadequate sampling effort. Species accumulation curves, which could have accounted for this limitation, were not generated.
ii) The sampling methods also do not account for detection issues (i.e., false negatives; Sollmann et al. 2013). This is especially pertinent given that a much higher number of mammal species occur in the region, which find either inconsistent, or no mention in the Railway Report (See Appendix III of this paper). For example, the Executive Summary (Page 5-6) mentions 42 mammal species were found using a literature survey, but the presence of the Bengal Tiger (India's National animal) is not explicitly stated. Appendix 2 of the Railway Report (Page 166) mentions 23 species of mammals, but does not mention which of those are Schedule I species, even though the region has 11 Schedule I mammal species. The ecological value of the region may have been underemphasized due to these inadequate methods as many more mammal species that occur in the region are likely to have been missed as they were not accounted for (Hayward et al. 2015).
iii) The description of the methods is very sparse and limits clear understanding (Page 153). The sampling unit was undefined -signs were recorded both inside and outside of belt transects. The study description lacks any detail about statistical methods used to assess species richness or percentage occurrence or relative abundance, using indirect signs or direct sightings. iv) Randomly placed belt-transects used in the Railway Report are not a suitable choice to assess large and small carnivore species richness and occurrence (Barea-Azcón et al. 2007). Further, signs were recorded opportunistically from outside of belt transects (Results, Page 153-154), but no clear analytical framework is provided for this data. Carnivores often tend to move on forest trails, roads, dry streams therefore a non-random or systematic sampling approach (within beats or grid cells) would be more appropriate to specifically assess J TT carnivore occurrence in the study region (Karanth et al. 2011).
v) Camera-traps are one of the best tools available to assess the occurrence, density, and abundance of mammals (O'Connell et al. 2011). But, the Railway Report uses a sparse sampling effort by surveying only 16 sites (camera-traps malfunctioned in nine of the 25 sites surveyed). In addition, the cameras were placed for less than six days in most sites. Studies have found a minimum of 20 to 30 error-free days of cameradeployment are required for stable estimates of species occurrence (Hamel et al. 2013). The standard duration for density assessment of large cats in Tiger Reserves and PAs of India is 25 days (with a closure period of 45-60 days). Therefore, a sampling duration of less than six days used in the Railway Report translates to poor data collection, which eventually affects any ecological inferences derived from such studies (Burton et al. 2015).
vi) The camera-trapping protocols lacks any detail about the camera models used, mode of deployment, camera-settings, and study design (Meek et al. 2014).
vii) Table 2.8.1 (Page 154) reports the species Viverra zibetha (Large Indian civet) which is not found in the Western Ghats, but in northeastern India. The table also mentions the occurrence of an otter species, Lutra lutra, the Eurasian Otter, which has not been recorded from the region. The Railway Report provides no evidence of its presence in the form of photographs. Two other species of otters which have been recorded and photographed in the region, the Asian Small-clawed Otter and the Smooth-coated Otter are not mentioned (Punjabi et al. 2014;Krupa et al. 2017). Page 161 of the Railway Report has erroneously labelled Wild Pig Sus scrofa as Indian Porcupine Hystrix indica.
viii) Appendix 2 in the Railway Report (Page 166) has incorrect coding for species: Langur and Bonnet Macaque are listed as herbivores (when they are actually primates); Asian Palm Civet is coded as a carnivore, but the Small Indian Civet, Brown Palm Civet, and Stripe-necked Mongoose are incorrectly coded as herbivores; the otter and Indian Pangolin are coded as large mammals, but the Asiatic Wild Dog, which is larger in size is coded as a small mammal. This reveals a naive understanding of mammals and the impacts that railway expansion could have on low-density species such as carnivores.

Land-use
i) The land-use land-cover map was derived from classification of single date satellite data, acquired in April 2013. Since the project area supports different types of vegetation which have variation in spectral signatures during different seasons, an ideal mapping exercise should have considered seasonal data, for at least two different seasons within one year.
ii) Out of six effective bands of Landsat and eight for vegetation discrimination, only four bands have been used for classification. This essentially leaves out the details of land-cover class categories that are clearly identified by the other two short-wave infra-red bands. These two short-wave IR bands demarcate the response of vegetation to moisture stress, and thus improve the classification of the forest types (Ferreira et al. 2016).
iii) The reasoning behind the number of sampling points used for each land-cover category is not clear. It is stated that unsupervised classification, which yielded 15 classes, was used as a basis for ordering the landscape into distinct units. It is unclear, however, if these 'distinct units' were further assigned land-cover classes on the basis of any reference map. A reference map could have informed the locations where ground truth data was necessary for ascertaining land-use types.
iv) The exact methodology for land-cover classification, parametric (maximum likelihood, minimum distance to means), or non-parametric (support vector machines or any other) has not been mentioned. This prohibits a nuanced understanding of the method of classification for a forest complex.
v) Ancillary data such as topographical information from an elevation model have not been utilized for assessments. A simple elevation profile of the proposed railway route indicates an elevation range of 80-500 m. In a high elevation area with varying gradients, the topography of the land determines much of the vegetation assemblages, and this could be important information to include in the classification process. The importance of topographic information for vegetation mapping is a widely accepted methodology (Das et al. 2015;Roy et al. 2015) and earlier work in the Eastern Ghats region has used topographic information effectively to this end (Balaguru et al. 2003).
vi) The basis for accuracy assessment has not been mentioned. An overall accuracy of 88% is indicated, but no reference map seems to have been used for calculation. The report also does not mention the percentage of samples used for training and testing the classification, which is a standard accuracy assessment procedure.

Water
i) Water pollution is a major concern during the construction as well as during the operation phase. Water pollution analysis, however, was minimal with no monitoring of pollutants done for polycyclic aromatic hydrocarbons (PAHs) and heavy metals because of the existing railway-line, despite high concentrations being often reported in waterways bisected, or bordered by railways (Wiłkomirski et al. 2011;Wiłkomirski et al. 2012;Levengood et al. 2015).
ii) Furthermore, Escherichia coli bacterial contamination was reported in all sampled streams, indicating faecal contamination, which may be attributed to waste disposal from passing trains. The total coliform count ranged from 221/100mL to 542/100mL, while the safe threshold value is 100 count/100mL. The increased risk of coliform contamination resulting from the railway expansion is a severe threat, as many streams that cross the tracks harbour sensitive wildlife, and also supply water to villages downstream for drinking and farming.

Air
i) No air quality monitoring was performed to provide baseline levels or to establish the risk of railway expansion in this region. The Railway Report assumes that engines will be electrified; however, if existing diesel engines are used then the doubling would increase the amount of pollutants associated with combustion and diesel emissions.
ii) The main constituents of diesel engine exhaust emissions are Carbon (CO, CO₂), Nitrogen (N), Nitrogen Oxides (NOx), Sulphur Oxides (SOx), Hydrocarbons (HC), Methane (CH4), Non-Methane Volatile Organic Compounds (NMVOC), PAHs, and particulate matter (PM) (Borda- de-Água et al. 2017). Monitoring of the current pollutant levels should have been performed at least twice a year to avoid data bias due to seasonal variation, although quarterly (or even monthly) sampling events could have been employed (Jayamurugan et al. 2013;Manju et al. 2018).

Soil
i) Chemical properties of soil and baseline levels of soil pollution were not established during sampling and analysis. Soil and plants surrounding the railway lines should be monitored for organic and inorganic compound contamination, resulting mostly from used lubricant oils and condenser fluids, the transportation of oil derivatives, metal ores and other chemicals, as well as from application of herbicides and other treatments to the train vehicles. These pollutants, however, were not considered in this assessment.
ii) PAHs, heavy metals, oil-derived HC, and to some extent, polychlorinated biphenyls (PCBs) should be monitored in soils, with risks comprehensively assessed as they exhibit toxicity, long-term stability and a cumulative effect in the environment (Wiłkomirski et al. 2011;Wiłkomirski et al. 2012;Levengood et al. 2015;Pereira et al. 2015). PAHs are carcinogenic and mutagenic to living organisms (IARC 1989). The main source of PAHs in railway areas are machine grease, fuel oils and transformers oils. Heavy metals (such as Pb, Cd, Cu, Zn, Hg, Fe, Co, Cr, Mo) originate mainly from material abrasion and fuel combustion in diesel and electric locomotives, therefore the railway expansion will lead to further heavy metal contamination in soils.

Social Impacts
i) The Railway Report's socio-economic survey of 60 families conducted in four villages does not report the total number of affected families, demography and livelihood patterns of concerned villages. The sampling strategy and the criteria for selection of households is unclear. The questionnaire was focussed on the perception of transport models by local communities. The questionnaire did not have open-ended, nonleading questions to bring out local concerns towards the project, and possible impacts on their livelihood and environment. Instead, it addressed questions such as preferred mode of transport, where 90% of the respondents listed trains.
ii) The Railway Report mentions a public consultation meeting regarding the railway expansion project that occurred in June 2016 at Kulem Panchayat (Hindi: Village Council) office (Page 190). The Kulem Panchayat raised concerns about the impact of the project on the Dudhsagar waterfall which contributes revenue from tourists to the local economy, availability of medicinal plants and disturbance to the temple close to Sonalium Station (Page 191). The consultation meeting was attended by only 14 members, most of whom were panchayat office bearers and members of the biodiversity committee, but not by the general public who would be affected by such developmental projects. As this meeting took place in 2016, before the Railway Report was published (in 2017), it is unclear whether a public hearing took place after the report was published. This suggests that the affected public is unaware of the damage the expansion may bring to their livelihoods.
iii) The Railway Report mentions that NTFPs and medicinal plants from the forest area were important for local use (Pages 169-171), but the specific impacts of J TT the railway expansion on such NTFP and medicinal plant species were not assessed. Datar & Lakshminarasimhan (2011) reported that local communities around BMWS were dependent on the forest for wild edible mushrooms, fruits, herbal medicinal plants, and specific plants for cultural use. This indicates that it is important to assess the impact of the proposed project on NTFP collection.
iv) The Railway Report finds that existing faecal contamination in the streams near to the railway tracks and the level of contamination is already 2-5 times the prescribed limit. Waste generation due to construction debris within the forest can further pollute soil and water resources in this sensitive region, thereby also affecting human communities. Increased waste dumping by railway passengers near villages can attract wildlife to these villages, which can result in human-wildlife conflict scenarios.

REVIEW OF THE 400kV TRANSMISSION LINE
The transmission line project did not have the assessment study in the public domain and therefore this limited our review to aspects of this project for which information was available in the public domain on the Parivesh portal. The key concerns with the transmission line project are discussed here.
i) The construction of new power lines in forest areas of high conservation value should be avoided (Eldegard et al. 2015). The transmission line project passes through a PA (11.54 ha inside PA) and the total forest land required for the project is 48.3 ha (almost 50 ha, for which an EIA is necessary from a socio-ecological point of view). The minutes of the meeting of the Goa State Board for Wildlife held on 02 December 2019 mentions that "the Biodiversity Impact Assessment studies and Biodiversity Management Plan has been prepared by ERM India Pvt. Ltd, Gurgaon has been submitted". The same, however, is not available in the public domain to allow a clear assessment of projected impacts.
ii) The detailed project report that is available for the transmission line makes contradictory statements about the location of the transmission line in the BMWS & NP. It first states that 2.51km of the transmission line is within the NP, clearing an area of 11.54ha (Table 1, Page 2, Detailed Project Report). Subsequently, when justifying the reason for choosing between alternative routes of the transmission line, it states that the chosen route fully avoids the NP. These statements severely weaken the report and hinder an effective assessment of the impacts of the transmission line, which already lacks sufficient public scrutiny. An inspection report by the forest department indicates that over 4,146 trees and 985 cane clumps in the PA are to be cut for the project.
iii) The project proponent claims that "transmission line projects are environment friendly and do not involve any disposal of solid effluents and hazardous substances in land, air and water. Moreover, forest area trees are felled below each conductor to facilitate stringing. On completion of construction only one strip is maintained for O & M purpose. Therefore, the actual loss of forest is restricted to some selected areas only." These statements do not recognize the larger effects of the transmission line on birds and volant mammals such as bats and gliding squirrels, or on arboreal species such as the Slender Loris, Giant Squirrel, Bonnet Macaque, and Grey Langur. For example, due to the absence of tree cover along transmission lines, arboreal mammals such as Lorises are forced to use electric wires of power lines to cross, causing mortality due to electrocution (Raman 2011).
iv) The project requires a clearance for 35 years, during which there will be regular cutting below the transmission line. This is especially concerning given that the project cuts through the PA, so the effects of this project are long-term.
v) The statement "the actual loss of forest is restricted to some selected areas only" fails to take into account existing evidence that power lines are linear intrusions that prevent animal movement, fragment communities of small mammals (Goosem & Marsh 1997), and cause mortality due to electrocution and collision (Jenkins et al. 2011;Rioux et al. 2013;Loss et al. 2014;Uddin 2017). Large mammals have also been electocuted due to sagging power lines (Raman 2011). The area underlying the proposed transmission line currently (i.e., without the construction of the power line) offers low resistance to large mammal movement, indicating that the area is important for animal movement (Jayadevan et al. 2020; https://indiaunderconstruction.com). In their paper, Jayadevan et al. (2020) recommend avoidance of new infrastructure in areas that currently pose a low resistance to movement.
vi) Transmission lines have several impacts on birds. Studies have shown that birds avoid areas between 0.25 and 0.6 km of transmission lines (Dunkin et al. 2009;Gillan et al. 2013). Transmission lines cause bird mortality due to electrocution and collision (Uddin 2017;Biasotto & Kindel 2018). For example, many birds use structures of transmission lines as a perch, which often leads to electrocution (Biasotto & Kindel 2018). The clearing of J TT trees for the transmission line affects the movement and nesting success of birds (Biasotto & Kindel 2018).
vii) The conservation value document uploaded by the wildlife warden details the damaging effects of the project. The document, however, concludes that the movement of faunal species will not be affected by the project, and the loss of trees can be compensated via afforestation. This is inaccurate, as transmission lines would impact movement of fauna, in addition to other deleterious impacts including mortality, as we detail above. Further, compensatory afforestation at a different site does not ameliorate any of the ecological impacts within the PA, as mentioned in the document.

DISCUSSION
We argue that mitigation measures proposed in the Road Report, Railway Report, and documents for the transmission line are inadequate and will not alleviate serious damage to the BMWS & NP or ecologicallysensitive regions around the PAs. We have explained this in detail in the following sections.

INADEQUACY OF MITIGATION MEASURES FOR NH-4A
i) For the mitigation measures, the Road Report merely notes that "Mitigation of man versus animal conflict is going to be the important issue that will threaten wildlife in Sanctuary area" (Page 97, Section 6.3.9 (i)). There are, however, no mitigation measures recommended to reduce the conflict created by road expansion. An acknowledgement of an important socio-economic and environmental problem will not equip the Goa Forest Department, National Highways Authority of India, or the Public Works Department of Goa to effectively manage the problem created by road expansion without detailed mitigation plans.
ii) For terrestrial fauna, the Road Report states that no impact on the wildlife is anticipated and hence does not outline any mitigation measures (Page 97, Section 6.3.9 (II)). Given that nearly 32ha of forest land will be diverted for the project, there is likely to be an impact on wildlife. There is growing scientific evidence demonstrating that building new roads and their upgradation or expansion has serious impacts on wildlife in protected areas. For example, Garriga et al. (2012) found a total of 2,013 wildlife mortalities on roads within protected areas of Catalonia, of which 267 were mammals (13.3%), 253 birds (12.6%), 245 reptiles (12.2%), and 1,248 amphibians (62.0%). A total of 85 different species were affected across all taxa due to roads within PAs over just two seasons, Spring and Autumn, in one year.
iii) As a measure to mitigate vegetation and habitat loss, the Road Report mentions that "an avenue plantation programme shall be promptly adopted to restore and further enrich the loss of vegetation" (Page 96, Section 6.3.9 (i)). Such measures may increase green cover, but they do not mitigate the impacts of road construction on vegetation or wildlife. Instead, it also puts people at risk due to the increased likelihood of vehicular collision with mammals (Case 1978;Jaren et al. 1991;Putman 1997;Cain et al. 2003).
iv) The Road Report proposes "periodic maintenance of drains to check scouring of soil" to decrease soil erosion (Page 92, Section 6.3.5). Soil erosion is expected to be higher in tropical forests, such as BMWS & NP, due to its wet climatic conditions and steep terrain (Sidle et al. 2006;Sidle & Zeigler 2012). Deposition of eroded soil into rivers at an increased rate is responsible for increasing turbidity and temperature of the water, reducing the amount of dissolved oxygen and changing existing flow regimes, while accelerating eutrophication (Beevers et al. 2012;Douven & Buurman 2013). The proposed clearing of land for the development of the road is likely to make cut sections highly susceptible to soil erosion. Drainage structures and culverts are essential to allow better above-ground water drainage, and prevent drastic changes to the hydrology of the landscape and decrease flooding along the road during monsoon seasons (Sidle et al. 2006;Laurance et al. 2009). No site-specific hydrological survey has been carried out to arrive at the optimal number of culverts and bridges, and their spatial placement.
v) Although the Road Report aims to reduce the impact of the developmental project in the "direct path" of the roadworks, it is pertinent to understand that the impacts of road construction are rarely limited to the direct path. Environmental impacts of roads extend beyond the direct impacts of construction and tree clearing, to indirect impacts because of increased human access and vehicular traffic. This includes, but is not limited to, air, water, and noise pollution, disturbance effects, fragmentation due to edge effects, and hindrances to migratory corridors (Alamgir et al. 2017).

THE IMPACT OF ROADS
We further expand on biotic and abiotic impacts of roads here, for which no mitigation measures have been suggested.

J TT i) Roads compound the impacts of natural disasters
Constructing roads in hilly and mountainous terrain increases the risk of natural disasters such as landslides and flooding (Sidle et al. 2006;Larsen & Torres-Sánchez 1997;Larsen & Parks 1998). There is no information on the susceptibility of the proposed site to extreme weather events in the EIA. Such dissemination of information regarding the socio-economic and environmental risks involved in the project is critical to the decision of investors, decision-makers and taxpayers, whose money is being utilized for the project. Road projects that pass through forested areas and lack proper planning can lead to major cost overruns, corruption, and damage to the environment (Trombulak & Frissell 2000;Alamgir et al. 2017)

ii) Roads are a cause for wildlife mortality (roadkills)
Enabled by the expansion of the highway, an increase in vehicular traffic in the area can be expected. This will likely increase the rates of wildlife-vehicle collisions, impacting species of most terrestrial fauna. A study from Mudumalai Tiger Reserve found road mortality of 40 animal species, including amphibians, reptiles, birds, and mammals (Baskaran & Boominathan 2010). Additionally, animals that are slow-moving or burrowing, such as freshwater turtles, amphibians, snakes, and soil-living fauna, get killed during road construction. The impacts of earthwork and annual maintenance operations on terrestrial fauna are usually overlooked (Clevenger et al. 2003;Fahrig et al. 1995;Trombulak & Frissell 2000;Goosem et al. 2010).

iii) Roads are barriers to wildlife movement, and cause habitat fragmentation
For many species, particularly in the Western Ghats, the expansion of the NH-4A is an additional fragmentation of an already fragmented habitat . The resistance to potential large mammal movement posed by the existing NH-4A is higher than the median resistance to mammal movement in the Western Ghats (Jayadevan et al. 2020; https:// indiaunderconstruction.com). Expansion of the road can, thus, lead to an increase in the resistance posed to movement, and lead to increased isolation between forest patches on either side of the road.
Subdivision of remnant forest patches due to various linear intrusions such as highways and roads causes "internal fragmentation" (Goosem 1997;Goosem 2007). Such internal fragmentation with wide, cleared roads and their edges, physical barriers such as fences and crash barriers, cuttings, fill batters, and culverts with drop structures, could be a serious threat to movement of wildlife and lead to increased negative human-wildlife interactions . For example, many animals in tropical forests avoid even narrow linear clearings (< 30m wide; Holderegger & Di Giulio 2010; Laurance et al. 2009). Increased traffic and continuous vehicular movement can stress the animals or make species alter their behaviour in the vicinity of roads (Trombulak & Frissell 2000). While certain species such as macaques are attracted to roads for scrap food from travellers (a potential ecological trap), species such as Elephants have been observed to avoid roads and highways due to associated risks, or suffer mortality from collisions (Blake et al. 2008). Behavioural avoidance of the road may also be exhibited by animals that can fly over the width of the road (e.g., birds and bats), due to the noise, pollution, and risk of crossing (Laurance et al. 2009).
The problem of fragmentation by roads is particularly acute for canopy dwelling species that use closedcanopy structures to move and do not generally use the ground to cross. In the absence of tree cover, treedwelling animals are forced to either use the ground or cross using power lines, which can lead to mortality due to vehicular collisions or electrocution. This is especially the case for primates, arboreal rodents, and some carnivores (Radhakrishna & Singh 2002;Raman 2011).

iv) Roads affect the genetic diversity of animals
Decreased movement of animals across roads leads to decreased genetic variation, due to reduced genetic exchange between populations. For example, studies from India show that roads negatively affect tiger connectivity (Joshi et al. 2013;Dutta et al. 2018;Thatte et al. 2018). Such impacts can be seen after just a few generations in populations of large mammals that have been separated by newly built roads and highways (Holderegger & Di Giulio 2010).

v) Roads affect biodiversity due to increased noise pollution
Although monitoring of noise quality levels created by the existing highway was carried out at eight sites designated as commercial, industrial and residential, there was no monitoring carried out on existing highway stretches within the protected area. Noise quality levels were found to be "within the limits" for commercial and industrial categories but "exceed the limits" in the residential category. Noise pollution associated with roads has been shown to decrease reproductive capacity in bird and amphibian species, as well as in mammals J TT such as Tigers (Kerley et al. 2002;Hoskin & Goosem 2010;Qin et al. 2014;Laurance 2015), with impacts seen at the community level as well (Francis et al. 2009;Slabbekoorn & Halfwerk 2009).

vi) Roads lead to increased human accessibility
Roads passing through forested areas increase human accessibility and can increase movement, settlement and human activity in frontier forest areas. This has manifold repercussions including forest fires, waste disposal and pollution, illegal timber harvest, poaching and hunting (Alamgir et al. 2017). Studies from protected areas in developing economies show that road expansion and improved accessibility to the market can result in expansion of agricultural and livestock frontiers with reduction in nearby forest areas of the protected area (Ratner et al. 2007;Lama & Job 2014;Phaipasith & Castella 2017;Walelign et al. 2019). Conversion from subsistence agriculture to cash crops, emergence of commercial service economies such as mass tourism resulted in transition from a low-impact economy to a high-impact one (Walelign et al. 2019). Local socioeconomic inequality also increased after road-expansion (Ratner et al. 2007). In the long run, the negative impact on the forest, waste generation and excessive use of agro-chemicals resulted in lesser availability of clean water, reduced soil fertility and local extinction of NTFP species (Phaipasith & Castella 2017). This also affected local governance systems negatively and people often could not revert to their subsistence economies which were relatively sustainable (Lama & Job 2014).

vii) Roads as a cause for habitat loss and degradation
During the construction and maintenance of roads and highways, habitat loss and degradation is observed due to direct clearing of vegetation, dumping of excavated earth and materials, regular usage of access roads by heavy machinery, and construction of labour camps. Within tropical forests, disturbance from roads due to fluctuations in light, temperature and humidity, increased mortality of trees beside roads, and spread of exotic species to a width of least 100m from the road (Laurance et al. 2009). Thus, "each kilometre of road directly and detrimentally affects at least 10 ha of habitat", and the impacts may persist for decades (Laurance et al. 2009;Raman 2011).

viii) Roads as corridors for invasive species
Roads have been found to be a major factor in the spread of invasive flora and fauna into forests (Mortensen et al. 2009;Meunier & Lavoie 2012). These invasive species can use the edge habitats along the road and invade forests by secondary wind dispersal, that would have otherwise been inaccessible (Kowarik & von der Lippe 2011).

THE IMPACT OF RAILWAY-LINE DOUBLING AND INADEQUACY OF MITIGATION i) Air quality
No potential impacts on air quality were studied, as the railway line between Castlerock and Kulem was assumed to be electric. If the trains in the proposed stretches run on traditional diesel engines, increased locomotive traffic due to the doubling of the railway line will lead to an increase in harmful exhaust components. The main pollutants from diesel locomotives are Carbon Dioxide (CO2), Carbon Monoxide (CO), Sulphur Dioxide (SO2), Nitrous Oxide (N2O), particulate matter (PM), hydrocarbons (HC), among others. Many of these pollutants are carcinogenic and responsible for health and environmental impacts (Lucas et al. 2017).
The report suggests monitoring of air quality and minimizing air pollution due to dust particles, vehicular and locomotive emissions, during the construction and operational phase. Although necessary, such general recommendations on controlling impacts on air quality during the construction phase will minimally help in reducing pollution as the project will take three years to construct and the operational impacts will be near permanent. Abrasion of brakes, wheels, dust, mineral transport will all still produce PM emissions even if electric locomotives are used (Levengood et al. 2015). No amount of mitigation will compensate for the longterm impacts of air pollution due to the proposed expansion.

ii) Sound (Noise pollution)
The noise levels at various regions within the areas of the proposed project were already noted to be above the permissible level of 91dB, posing a serious threat from noise pollution. Anthropogenic noise can affect acoustic communication among bird species that use calls and songs for a variety of functions such as attracting mates and defending territories (Collins 2004;Marler 2004). Noise emission from railways has also been documented to reduce the density and nesting behaviour of birds, with nests that are farther away from railway lines being more successful (Mundahl et al. 2013).
To reduce noise pollution, the Railway Report recommends switching to electric engines, planting native tree species along the railway line and building sound barriers on both sides of the track, particularly J TT within the biodiversity-rich Kali Tiger Reserve (from Castlerock station -Goa border) and BMWS & NP. All of these three recommendations, however, have serious drawbacks which the report has overlooked. a. Switching to electric engines will not greatly reduce noise pollution. This is because, at a speed of 30-200 km/h (the speed at which most trains will be travelling between stations), all trains, including the electric trains, produce a "rolling noise" which is the dominant source of noise pollution (Clausen et al. 2012). b. Recovering vegetation beside roads and railways is known to attract wildlife and increase their vulnerability to get killed by moving vehicles (Case 1978;Jaren et al. 1991;Putman 1997;Cain et al. 2003). c. Sound barriers will further intensify the impacts of forest fragmentation caused due to railways. While the size and structure of sound barriers are not mentioned in the report, they are usually artificially built, vertical walls of a solid structure, which blocks the noise created by moving trains. While this could be a reasonable mitigation measure for railway lines that pass through human habitat, it will have detrimental ecological impacts inside a forest ecosystem. Sound barriers can cause various negative impacts on wildlife, particularly through isolation of populations (Bank et al. 2002). Given that the minimum height of such barriers is as high as the train, and with electric lines proposed to be running on top of the railway, it would make it impossible for any terrestrial species to freely move to the other side of the track and will be a death trap for wildlife trapped between the barriers.

iii) Water quality
The Railway Report states that the current water quality in streams along existing railway track is pristine (Bureau of Indian Standards 2012), but with harmful levels of bacteria Escherichia coli in all streams (221 to 542 per 100mL), it indicates widespread faecal contamination of waterbodies mostly due to exisiting train traffic (threshold is 100 per 100mL). Creation of new railway embankments for the proposed double gauging will further lead to vegetation loss, soil compression and changes in water drainage, thus increasing runoff, promoting erosion of topsoil and increasing water turbidity (Ferrell & Lautala 2010;Chen et al. 2015). Turbid water has been found to affect the diversity and abundance of aquatic wildlife communities such as odonates and freshwater fish (Luce & Mountain 2002).
The Railway Report mentions that the new railway coaches will be fitted with bio-toilets, hence reducing the likelihood of E. coli infiltrating streams along the railway route. Construction of dykes and retaining walls along the railway line to restrict the movement of sediments during the construction phase has been recommended. While this may address sediment runoff, it may indirectly inhibit animal movement, adding to the effects of tree clearing, noise, and train movement.

iv) Biodiversity
The major focus of the studies appears to be to create baseline information on species diversity and abundance, rather than to explicitly study the impact of the proposed expansion on biodiversity. The Railway Report only cursorily mentions that animal movement will be impacted by the doubling of the railway line (Chapter 18, Page 207) and does not address long-term impacts to landscape connectivity that all the taxa under study face from the proposed expansion of the railway line. There is strong evidence of the negative impacts of railway lines on biodiversity.
Railway lines have been shown to be barriers to movement for large mammals such as the tiger (Dutta et al. 2018). The current railway line between Kulem and Castlerock poses a high degree of resistance to large mammal movement (Jayadevan et al. 2020). Doubling of the railway track will lead to a higher frequency of trains, and further increase resistance to movement. This can isolate the forest patches on either side of the railway line. In addition to its impacts on movement, noise and vibrations from railways affect insects, amphibians, and birds. Further, the availability of food (solid food waste; carcases of dead animals) and vegetation along the railway edges attracts reptiles, few species of birds, and several mammals acting as an ecological trap and leading to higher mortality due to collision with trains (Lucas et al. 2017).
The mitigation measures suggested in the Railway Report are very general. The suggestion of the creation of 'biodiversity parks' for conserving birds and mammals is not compensatory, when the protected area, which is a biodiversity-rich region, will be fragmented. For aquatic life, it is suggested that railways should adopt the 'best construction procedures' to reduce turbidity, siltation, etc., but what these procedures comprise of is unexplained.
The Railway Report highlights cases of Gaur and Sambar being hit by trains on the existing single track, reaffirming that the doubling of the railway line will lead to increased risk of accidental collisions with wildlife.

J TT
Although the report identifies 42 animal-crossing points for mammals, a bare minimum of four animal underpasses are finalized at Ch 32/200, Ch 41/100, Ch 45/500, Ch 49/500 (RVNL Letter No.PlU/UBLllLN654 dated 06.10.2018 to deputy forest officer). The report suggests many other mitigation measures to be followed (Pages 207-208), but such mitigation measures are undetailed, and without strong supervision during implementation have poor application in practice.

INADEQUATE MITIGATION MEASURES FOR THE TRANSMISSION LINE
i) On the subject of mitigation measures for the transmission line, the inspection report of the transmission line mentions only that "the user agency has agreed to cut minimum trees requirements and to adopt wildlife-friendly mitigation measures." It adds that "trees listed for felling under this project will be compensated in the long term through the proposed compensatory afforestation programme covering double the degraded forest land." It is not clear how 'minimum tree cutting' will be calculated or enforced. No details are provided on where and when the compensatory afforestation will be executed. Further, without an impact assessment of the transmission line, it is not clear what 'wildlife-friendly mitigation measures' will be implemented.
ii) The inspection report fails to take into account the ecological impacts of the transmission line as we have detailed in this paper (Section 5). A background paper for the National Board for Wildlife (Raman 2011) recommends that the first priority for power lines in forests should be prevention, followed by re-alignment. The third option of a mitigation measure is suggested only where the first two have been comprehensively considered and ruled out with sufficient justification (Raman 2011). In case a transmission line passes through a biodiverse region, recommended mitigation measures for transmission lines include insulators on wires to avoid bird electrocution, placing of perch deterrents on cross-arms and poles and using large linemarkers on earth wires to increase their visibility during the day and night, thus avoiding collisions by birds and volant mammals (WII 2016). But neither of these are considered as mitigation measures for this project.

CONCLUSION
Any major infrastructure projects should be avoided within PAs, unless there are exceptional circumstances that will clearly show forest diversion will benefit wildlife (as per the WPA, 1972). Utmost importance should be given to all environmental and ecological impacts of any project, and as per the background paper of the National Board for Wildlife itself, ecologically-harmful projects should be avoided. In the present case, there is not one, but three large projects which are planned in this ecologically-sensitive region. It is noteworthy that the Western Ghats is a designated Natural World Heritage Site by UNESCO. The cumulative impacts of these three projects may change the entire ecology of the BWWS & NP, as well as the neighbouring Kali Tiger Reserve, and will result in irreparable damage to its fragile environment. Further, such damage will impact the quality of human life within and near the PA. Multiple projects also call for an in-depth investigation into cumulative impacts on the PA. Cumulative impact studies have been considered mandatory in many countries (Braid et al. 1985), and are implemented rigorously for their added value in understanding irreversible changes to existing natural systems (Xue et al. 2004).
It is pertinent to note that two of these projects (NH-4A and transmission line) were awarded wildlife clearances in the 57 th meeting of the Standing Committee of the NBWL, held on 7 April 2020 through a video conference, which is unlikely to have had critical evaluation. Our review details how the EIAs and assessments for these projects are considerably weak, and evidently overlooked by the highest statutory authority that is mandated to protect wildlife in the country. Sociallyand environmentally-just development is important, but none of these projects provide any benefit to wildlife or the environment in the BMWS & NP. Environmental costs and mitigation measures are not comprehensively assessed in the EIAs and assessment studies. Information on the land area for compensation, overseeing agencies for mitigation measures, monitoring and penalties for non-compliance are also not laid out in detail.
Faulty EIAs and other assessment studies continue to be condoned by successive appraisal boards and governments, with a lack of due process. Such practices consider environmental concerns as a burden on development, rather than a process that guides sustainable development, which should, therefore, be strengthened. This further weakens socio-ecological governance in a country which is ranked a 168 (out of 180 countries) in the Environmental Performance Index

Environment impact assessments for projects in Bhagwan Mahavir WS & NP
Punjabi et al. (Wendling et al. 2020). Considerable opportunities exist to improve the EIA and assessment process in India (Paliwal 2006). Incentivising post-clearance monitoring and evaluation is vital (Duflo et al. 2013); however, a rational screening process which fortifies existing legislation and avoids forest diversion proposals in protected areas at the outset itself is most necessary (Rajaram & Das 2011).