Successive Cyclones Attacked the World’s Largest Mangrove Forest Located in the Bay of Bengal under Pandemic

Abstract

Despite the global focus on the COVID-19 pandemic, the promise of impact to tropical coastlines and stochasticity of destruction caused by tropical cyclones remains unaltered, forcing human societies to adapt to new unadaptable scenarios. Super Cyclone Amphan’s landfall—the third cyclone of the season within the world’s largest mangrove forest—brought a new uncertainty to this undeveloped region of South Asia. How do vulnerable people deal with multiple disasters that limit necessary humanitarian response while still maintaining the natural environmental integrity of a system harboring critical wildlife populations and protecting people from further disaster? We explored this reality for the Sundarbans region using a remote sensing technique and found that the western part of Sundarbans mangroves was severely damaged by Amphan, suggesting that rapid remote sensing techniques can help direct resources, and recognize the eventuality that response will be a best effort for now. If 2020 is a window, multiple disaster management scenarios may become more common in the future. Yet, society’s obligation for maintaining environmental integrity remains unchanged.

1. Introduction

Although a positive feedback between global warming and the degree of variability of tropical cyclone frequency and intensity has been a subject of debate, consistencies among simulation models and physical reasoning are capable of projecting at least 3 future scenarios: (1) decreases or no change in tropical cyclone frequency [1,2,3], (2) increases in intensity with clustering patterns and fractional increases in the number of more intense storms, and (3) increases in tropical cyclone-related rainfall rates [1,2,3]. Scenarios 2 and 3 explicitly apply to the northern Indian Ocean (NIO) where, despite reduced frequencies of cyclones compared to other basins, the destructive power of cyclones is often much stronger because of the lower social and economic development of the human landscape [4].

The Bay of Bengal (BoB), located at the eastern side of the NIO, is the largest bay in the world, and one of the most populated coastal settings (~500 million people) where the

Frequency of intense cyclones has risen recently [5], and with some of the deadliest reported cyclones in recent history. An analysis of all tropical cyclones that formed from 1982 to 2020 over the NIO indicated that the BoB was also a prominent source of cyclone intensification in comparison to the western Arabian coast. The BoB has a warm oceanic-sized water body throughout the year with present sea surface temperatures (SST) of 28–29 °C [6]. These SSTs are greater than the threshold SST for cyclogenesis (25.5–26.5 °C) [7], ensuring rapid intensification of any cyclone before it strikes land. Geometry of BoB is a “textbook example” of how to propagate a super cyclone to have maximum destructive influence. High probability of cyclone intensification, shallow-water concave bay, low flat coastal terrain, and the funnel shape of BoB are conducive for water propagation by strong winds and surge energy that become further concentrated through funneling as storms move toward land [8]. The most recent large cyclone was Super Cyclone Amphan, which made landfall accompanying heavy rainfall in the coastal areas of India and Bangladesh on 20 May 2020. This month was also recorded as the warmest in the region since 1880, as per NOAA, NASA, and WMO [9,10,11]. Importantly, Super Cyclone Amphan became the first super cyclone in recent history to impact an under-developed area of the world during a global pandemic. Additionally, after Super Cyclone Amphan, the BoB was hit by two successive cyclones, BOB02 (11 October 2020) and BOB (20 October 2020). Though impact severity of the two successive cyclones was less compared to Amphan, BOB02 generated heavy rain at landfall affecting the eastern coast of India (Andhra Pradesh) [12].

Amphan’s impact occurred at a time when lockdown due to the COVID-19 pandemic was a pervasive global priority affecting humanitarian response and forcing millions of local people to be evacuated and huddled in safe-shelters while attempting to socially distance. Globally, Amphan was among the deadliest natural disasters registered during the pandemic [Category 3 or H3 as per Saffir–Simpson Hurricane Wind Scale (SSHWS), 178–208 km h⁻¹; more than 100 fatalities] [12], which was compounded by the pandemic’s grasp on society’s ability to provide aid and an uncertain toll yet to come from group spread of the disease as neighbors, friends, and outside personnel provided necessary aid. Yet, Amphan was the third successive cyclone in a season that developed in the BoB, causing severe damage to the eastern part of the BoB coast. The other two were Fani in April/May and Bulbul in November of 2019 (H4 and H2, respectively) [12]; a cyclonic succession that is even more difficult to recover from as a new set of rules began to govern how people can interact.

Cyclone impacts to the BoB result in saline water surge rapidly killing freshwater fish used for subsistence and further affecting drinking water sources, as well as preventing some land uses to cultivation for up to five years. Coastal communities often mitigate this multi-year impact by moving to cities until they can return, a process hindered significantly by COVID-19. After the cyclone, infections (especially diarrhea) occur quickly, and on-going medical treatments were impacted. Recently, the same occurred in The Philippines where Super Cyclone Goni, recorded as the strongest cyclone in the last four years, hit Catanduanes Island on 26 October 2020, when many parts of the country were under pandemic lockdown orders as well. Before Goni, The Philippines was struck by 12 cyclones (including category 3 and 4) in the same pandemic year. Following the alarming forecast of Goni, around four million people were evacuated, similar to the situation of Amphan, raising concerns over reduced social distancing and further viral spread. Like the BoB, it is noteworthy that The Philippines holds large mangrove areas (2675.2 km²) occupying 32.8% of their coastline. Cyclone impact is also common. For instance, a large section of mangrove forest was devastated in the central Philippines due to Super Cyclone Haiyan (2 November 2013) that caused damage of $12–15 billion (USD) and >10,000 human casualties [13]. In normal years, such impact is very difficult to overcome; the pandemic made recovery much more difficult to manage. Mangroves likely did play an important role in protecting coastal peoples from wind and waves. However, increased frequencies of tropical cyclones and the potential for intensified Super Cyclones (e.g., Amphan and Goni) are particularly concerning with or without mangroves, especially as coastal populations increase and pandemic spread probabilities intensify.

The land-ocean interface of the BoB has global uniqueness because of the presence of the Sundarbans, the world’s largest contiguous mangrove forest located on the coast of two adjacent countries, Bangladesh (62%) and India (38%) [14]. The Sundarbans also draw people to this low-lying area, creating a conflict between societal provisioning and disaster vulnerability. Functioning of mangroves as bio-shields against natural disasters, like tsunamis and tropical cyclones, has the potential to save thousands of inhabitants at the coastal rim, but even large mangrove areas alone do not provide all the protections humans would like [15], though the coastal protection can be considered as one of the most undervalued ecosystem services of mangroves [16]. While our understanding of surge protection value is limited versus other ecosystem services provided by mangroves, such as protecting infrastructure, provisioning of wildlife habitat and carbon sequestration, mangroves are likely to be much more effective in normal wave suppression than they are surge protection [17]. Yet, going largely unnoticed was the statistically significant reduction in human deaths during a super cyclone landfall along the eastern BoB, India in October 1999 [18], or the protective role of coastal vegetation during the Asian tsunami in 2004 [19]. Mangroves certainly help to spare lives in multiple ways during storms; an influence that may be reduced with repetitive future cyclonic events.

The Sundarbans have global significance as a UNESCO World Heritage Site, registering large rates of carbon uptake from atmospheric CO₂ [20], providing for fisheries, and supporting a unique wildlife resource (e.g., Royal Bengal Tiger, Chital Deer, unique dolphins) while partially adjusting naturally to current rates of sea level rise [21]. At stake are protective or recovery measures to ameliorate human impacts to mangroves (e.g., as people gather additional sustaining resource from the mangroves) and wildlife impact to mangroves (e.g., destruction of a fence to keep tigers away from humans), which are compromised by pandemic rules that are still quite foreign to the Sundarbans residents. It is expected that this juxtaposition between normal response to a super cyclone and COVID-19 could worsen Amphan’s impact to the Sundarbans, and future cyclones, perhaps turning parts of the Sundarbans into a carbon source from delayed re-engineering or re-planting efforts requiring lots of people to gather in small spaces. In the Sundarbans mangroves, it has been reported that mangrove rehabilitation programs are one of the major determinants of turnover between mangrove and non-mangrove as well as encroachment, erosion and aggradation [22]. Reforestation and afforestation have been conducted in the Indian Sundarbans since 1989 to conserve endangered species such as Heritiera fomes and to establish bio-shields against cyclones by planting mangroves [23]. The moving control under pandemic could bring a delay of reforestation after the storm disturbance. The aftermath of the largest mangrove dieback in Australia yielded something similar vis-à-vis massive carbon loss [24]. Sediment erosion and inundation often result in accelerated forest land subsidence [25] with slow but steady submergence of many outer shore islands around the BoB. As site access is impossible after many cyclones, it is necessary to estimate the extent of damage reliably with remote sensing, identify potential eco-physiological disorders underway with the mangroves immediately after a cyclone strike, and begin to plan actions to facilitate mangrove and wildlife habitat recovery post-cyclone with due diligence once post-pandemic access returns. This study presents the results of monitoring the damage caused by a super cyclone Amphan using remote sensing technology that enables us to conduct a quick assessment, and proposes an approach for assessing a massively disturbed natural environment under moving control conditions, such as during a pandemic of COVID19, to facilitate environmental and cultural sustainability.

2. Materials and Methods

2.1. Study Area

The Sundarbans mangrove forest region (Figure 1), both Bangladesh and Indian part bounded by 21°32′–22°40′ N latitude and 88°05′–89°51′ E longitude, was the focus of the study. Comprised of numerous complex, interconnected tidal creeks or channels the forests receive continuous supply of freshwater. The Sundarbans mangrove forest is a unique ecosystem that supports a large number of flora and fauna. Due to its immense importance, the forest is heavily protected by law as part of UNESCO world heritage site both in India and Bangladesh [14,20].

About two thousand animal species including the Royal Bengal Tiger dwells in this wonderful ecosystem. It hosts over 300 plant species, of which over 27 are mangrove tree species [14,26]. Major mangrove species that dominate the forests are: H. fomes, Excoecaria agallocha, Ceriops decandra and Sonneratia apetala. However, patches and mixes of Avicennia, Xylocarpus, Bruguiera, and Rhizophora trees and Nypa palm exists throughout the forest [14,26,27]. The mean monthly temperature, rainfall and relative humidity varies between ranges from 12 to 35 °C, 1600–2000 mm and 70–80%, respectively [14,26]. The monsoon season, months of June to October, brings about 80% of the total rainfall [28]. Located at the tip of the Bay of Bengal, the Sundarbans mangrove forest is often impacted by frequent cyclonic storms of varied intensities (Figure 1).

2.2. Cyclone Data

Data on major cyclones that formed in the Northern India (NI) basin since 1982 were retrieved from the India Meteorological Department (IMD) [12] (Figure 1). Cyclones that impacted the Sundarbans or made landfall there since 2016 (i.e., a centroid point: 21°56′21.31″ N, 89°10′56.01″ E) within a radius of 370.4 km (200 nautical miles) were cataloged according to the methodology described in Mandal and Hosaka [2], with adjustments due to image availability. As multiple cyclones impacted the Sundarbans, only the strongest were selected each year. As a result, cyclones Mora in 2017, Titli in 2018, Fani in 2019 and Amphan in 2020 were identified as our focal storms (Figure 2). The India Meteorological Department (IMD) uses a slightly elaborated categorization of cyclones based on estimated maximum sustained wind speed (km h⁻¹; hereafter “wind speed”), but for comparison, the SSHWS was used for categorization (hereafter “category”). Minimum distance (km; hereafter “distance”) of cyclone eye wall track from the center of the Sundarbans was determined. Study-selected cyclones along with those identified previously by Mandal and Hosaka [2], as well as specific cyclone-defining parameters, are presented (

Table 1. Major tropical cyclones that impacted the Sundarbans since the year 1988 from Mandal and Hosaka [2] in comparison to the present study (Mora, Titli, Fani, Amphan). The Saffir–Simpson Hurricane Wind Scale (SSHWS) was used for categorization of cyclone intensity (category), minimum distance (km; distance) of cyclone eye track from center (i.e., a centroid point: 21°56′21.31″ N, 89°10′56.01″ E), and maximum sustained wind speed (km h⁻¹; wind speed).

Code	Name	Date	Category	Distance (km)	Wind Speed (km h−1)	Detected Area km2 (% of Damage)
C88	04B	28–29 November 1988	H3	24	204	1093 (20.4)
C89	Unnamed	23–26 May 1989	TS	200	102	25 (0.5)
C90	BOB 09/04B	17–19 December 1990	TS	210	83	24(0.5)
C91	Gorky	29–30 April 1991	H4	160	241	432 (8.1)
C92	Unnamed	22–24 September 1992	TD	33	56	93 (1.7)
C94	Unnamed	2 May 1994	H3	328	204	418 (7.8)
C95	Unnamed	24–25 November 1995	H2	255	157	659 (12.3)
C96	Unnamed	28 October 1996	TD	34	74	61 (1.1)
C97	Unnamed	18–19 May 1997	H1	243	120	281 (5.3)
C98	Unnamed	18–20 May 1998	H1	210	130	166 (3.1)
C99	Unnamed	8–11 June 1999	TS	123	65	336 (6.3)
C00	Unnamed	27–28 October 2000	TS	52	65	78 (1.5)
C02	Unnamed	12 November 2002	TS	94	102	29 (0.6)
C04	Unnamed	15–18 May 2004	H1	303	120	300 (5.6)
C05	Unnamed	1–3 October 2005	TS	115	74	279 (5.2)
C06	Unnamed	30 June–5 July 2007	TS	160	65	441 (8.2)
C07	Sidr	15–16 November 2007	H5	59	259	1291 (24.1)
C08	Rashmi	25–26 October 2008	TS	59	83	258 (4.8)
C09	Aila	25–26 May 2009	H2	122	120	53 (1)
C15	Komen	26 June–2 August 2015	TS	116	46	137 (2.6)
C16	Roanu	20–21 May 2016	TS	106	116	152 (2.9)
C17	Mora	28–30 May 2017	TS	281	111	17.9 (0.3)
C18	Titli	8–12 October 2018	H1	10	148	31.2 (0.5)
C19	Fani	26 April–4 May 2019	H4	146	213	66 (1.1)
C20	Amphan	16–21 May 2020	H4	81	241	550.1 (11.5)

TD tropical depression TS tropical storm, H1–H5 Category 1–5 cyclone based on SSHWS.

).

2.3. Multi-Spectral Remote Sensing Data Analyses

Landsat 8 Operational Land Imager and Thermal Infrared Sensor (OLI/TIRS) surface reflectance data (Earth Engine ID: LANDSAT/LC08/C01/T1_SR) were accessed using the Google Earth Engine (GEE) platform [30]. Provided images from GEE were already atmospherically corrected, and filtered of cloud and shadow masking using a custom made function in GEE based on pixel_qa band. We chose the GEE platform because it is widely accessible to many natural resource agencies.

We developed a vegetation mask using a supervised classification and regression tree (CART) method. Landsat 8 surface reflectance images from 2018 (total 77 images) were sorted and filtered in GEE, and Normalized Difference Vegetation Index (NDVI) was calculated for each image. NDVI was calculated as follows:

NDVI = (NIR − RED)/(NIR + RED)

(1)

where, NIR represents band 5 (B5) and RED represents band 4 (B4) data from Landsat 8′s Operational Land Imager. The first seven bands (B1 to B7) and the NDVI band were used as prediction bands. Four classifications were used to define the Sundarbans mangroves remotely; (1) vegetation class: mangrove forest vegetation; (2) water class: river channels and open water areas; (3) mudflat: bare soil and flooded areas associated with channels; and (4) bare soil: bare land and sand beaches (not classified as mud). For data training, randomly selected points (about 30–100 per class) were generated using Google Earth Pro and published maps. 70% of the points were used to train the classifier and the remainder of the dataset (30%) was used to assess accuracy using a confusion matrix [31]. A forest cover vegetation mask was then produced from a classified map, and further analyses of forest disturbances were conducted for vegetated areas only. Results of the supervised classification and accuracy assessment are summarized (

Table 2. Results of supervised classification and accuracy assessment.

No	Class	Area (km2)	Producer’s Accuracy	Consumer’s Accuracy
1	Vegetation	6077.37	0.99	0.78
2	Water	3855.14	0.88	0.96
3	Mudflat	624.16	0.71	0.83
4	Bare soil	117.80	0.54	0.86
	Total area	10,674.47
	Overall accuracy		0.86%
	Kappa coefficient		0.79

).

The pre-disturbance period from Landsat imagery was selected as January to March for all cyclones, which consistently represented periods prior to cyclone impact. This period allowed us to avoid cloud contamination and reduce seasonal phenological differences from vegetation. The post-disturbance period from Landsat imagery was selected from the same months, but from the year following impact for cyclones Mora (2017), Titli (2018), and Fani (2019). For Super Cyclone Amphan, pre-storm and post-storm images were collected from 6 April to 17 May 2020, and 24 May to 12 August 2020, respectively.

Different approaches have been used previously for assessing cyclone disturbances. These included spectral mixture analysis [32,33] as well as spectral index analysis based on such components as enhanced vegetation index (EVI) [34] and normalized difference moisture index (NDMI) [24]. However, in the absence of field data (especially during pandemic lockdowns) and for the purposes of our quick assessment needs, we used remotely sensed NDVI only.

NDVI is perhaps the most popular index used globally, because this index has strong correlations with leaf area indices [35]. We calculated NDVI for all pre- and post-disturbance periods, then we produced maximum value composites using GEE’s “quality mosaic” function. Post-disturbance NDVI composites were subtracted from pre-disturbance NDVI images, and mean values (μ) were analyzed. We carefully checked the resultant μ values for normality and calculated a selection threshold using the formula.

Threshold = μ ± nσ

(2)

where σ is the standard deviation of the mean difference in composite values and n is non-negative integer numbers. We used ±3σ (thus, n = 3) for cyclones Mora (2017), Titli (2018) and Fani (2019), and ±1σ (thus, n= 1) for Super Cyclone Amphan (2020). The basic assumption was that any change in NDVI values less than μ ± nσ (i.e., approximately 0.1) might reflect natural forest dynamics and not actual damage from cyclone disturbance. Therefore, those values of “n” were censored. Without field data, it was not possible to quantitatively assess the true threshold value; however, we compared our results for Super Cyclone Amphan (2020) with local news, expert opinion, and field photos which was helpful for verifying our rapid analysis. Similar change detection methodology has been applied to assess cyclone disturbances in the Everglades mangrove forests [36,37], and cyclone induced land use and land cover changes in central regions of Mozambique [38].

2.4. Empirical Models to Validate the Impacts of Cyclones on NDVI Change

To interpret the spatial pattern of NDVI loss (NDVI_loss: dimensionless) after Super Cyclone Amphan, the relationships of NDVI_loss to the distance from the trajectory of Amphan (Track: m), distance from the water’s edge (Water: m), and NDVI before cyclone attack (preNDVI: dimensionless) were analyzed with a general additive model (GAM) by applying the penalized thin plate regression spline function of latitude (X: m) and longitude (Y: m). In the GAM, X and Y values are treated as the random effect to avoid Type I error caused by spatial auto-correlation among neighboring pixels. We used the function, gam, from package, mgcv, available in R software ver. 4.0.2 [39]. To calculate Track and Water, we used a plugin, NNJoin, from an open-source GIS software program QGIS ver. 3.10.1-A Coruña [40].

3. Results and Discussion

3.1. Rapid Detection of Mangrove Damage by Remote Sensing

Post-Amphan, true loss or damage of the Sundarbans mangrove vegetation is unknown; the pandemic still governs access. However, some damage like many broken embankments, inland flooding, and uprooted trees are reported on electronic media and obvious in our remote sensing work. A large portion of the mangroves lining several creeks and waterways are turning brownish red (as burnt or scalded) after Amphan; a strong remotely-sensed indicator of wide-spread mangrove mortality. Likewise, in early July of 2020, Calcutta University scientists surveyed some of the affected islands of the Indian Sundarbans and observed widespread mangrove mortality resulting from a mix of storm winds and pervasive post-Amphan hypersaline flooding beyond physiological tolerances of mangroves (Figure 3).

Changes in NDVI before versus after cyclones that occurred between 1990 and 2016 in the Sundarbans of both India and Bangladesh were recently analyzed [2,28], see Table 1. Category 3, or higher-intensity cyclones that occurred in succession after 7 to 12 years destructively affected 20% or more of the Sundarbans mangrove forest area, offering a grim insight into what a future with more cyclones might mean when impact is compounded through successive cyclones. In general, the first cyclone strike can bring massive damages to trees and subsequent cyclones immediately after the first one bring less damages to trees compared to the first strike since vulnerable trees are already excluded by the first strike [41,42]. The areas of four major cyclones impacting the Sundarbans in successive years of 2017 to 2020 were analyzed using Landsat images, and all areas hit by cyclones in the Sundarbans between 1990 to 2020 or made landfall within a radius of 370.4 km (200 nautical miles) were analyzed over successive years since impact (Table 1). From this, the intensity of Amphan (C20) was the third most severe after Sidr (C07) and an unnamed cyclone (C95) over the last 3 decades, with Amphan causing 11.5% damage to the mangrove vegetation. In comparison, previous cyclones caused more damage, with C07 and C95 impacting the Sundarbans in 2007 and 1995, and registered 24.1% and 12.3% mangrove damage, respectively.

3.2. Empirical Models of NDVI Changes

The GAM analysis detected significant negative effects of Track, Water and preNDVI to NDVI_loss on the mangroves (

Table 3. Results of the general additive model (GAM) analysis for the relationship of altered NDVI after Super Cyclone Amphan (NDVI_alt: dimensionless) to distance from trajectory of Amphan (Track: m), distance from water edge (Water: m) and NDVI before Amphan (preNDVI: dimensionless). R² was 0.16. The smoothing term based on a thin-plate spline function of latitude (X in m) and longitude (Y in m) was treated as a random effect (d.f. = 29, GCV = 0.004724, p < 0.001).

Coefficients	Mean	(95% CI)	p
Intercept	−9.82 × 10−2	(−9.87 × 10−2 to −9.76 × 10−2)	<0.001
Track	7.40 × 10−3	(7.37 × 10−2 to 7.43 × 10−3)	<0.001
Water	7.02 × 10−5	(3.93 × 10−5 to 10.10 × 10−5)	<0.01
preNDVI	1.91 × 10−1	(0.19 to 0.20)	<0.001

). The smoothing term based on X and Y assigned a random effect and was significant (p < 0.001). The Sundarbans mangrove forest showed greater loss of NDVI from Amphan versus the Bangladesh part. However, the effect of Water was visually unclear even at the local scale (Figure 4b). This is because preNDVI tends to be smaller near the shore and showed a more significant negative effect to NDVI_loss than Water. As a result, the lower NDVI_loss near the shore can be visually confirmed because of the effect of lower preNDVI near the shore. This is a reasonable result since the vegetation near the shore is regularly disturbed by wind and waves which determine the local limit of natural mangrove habitat occurrence and should result in the lower preNDVI. Interestingly, a photo (Figure 3b) captured the general trend that we noted in the mangroves near the shore consisting of shorter individuals with green canopies. Taller individual mangrove trees in interior areas showed damaged canopies, as indicated by reddish hues or complete defoliation. Similarly, a previous study on the influence of Super Typhoon Haiyan on mangroves in the Philippines reported that mangrove trees taller than 10 m were much more severely damaged than shorter trees [43]. Tree size-distributed cyclone wind damage is a common theme in mangrove forests globally [18].

The main trajectory of Amphan was over the extreme western part of the Sundarbans. As a result, the extent of mangrove structural change was more prominent on the Indian side of the Sundarbans versus Bangladesh’s mangroves (Figure 5), and the statistical results based on general additive modeling also revealed that areas closer to the trajectory of Amphan showed significantly greater decreases in NDVI (p < 0.01) (Table 3). Recovery trajectories of cyclone affected mangrove ecosystems after multiple cyclone strikes is unknown but being monitored by remote sensing in the Sundarbans. On the basis of the previous reports related to recovery process of mangroves after storm disturbances, we can expect the Sundarbans mangrove forests to show quick recovery after storm disturbance [13,43,44], which has been observed in our image monitoring after the Amphan strike (data is not shown). This quick recovery of mangroves is because the mangrove forests have distinct disturbance adaptive traits. That is, some dominant mangrove species, such as Heritiera fomes, Excoecaria agallocha, Avicennia spp., and Sonneratia spp. have the ability to recover by sprouting, as reported in Myanmar [44] and The Philippines [43], though high mortality of Rhizophora spp. is notable [43,44]. In the case of Sundarbans mangroves, the dominant species are H. fomes, E. agallocha, Ceriops decandra and Sonneratia apetala [26,27]. Thus, three of four dominant species (H. fomes, E. agallocha and S. apetala) in Sundarbans can be considered as resilient mangrove species. However, beyond the ecological resilience of mangroves, we should pay attention to a possible delay of recovery of mangroves owing to moving control caused by COVID19 pandemic and delayed human intervention.

3.3. Importance of Timely Assessment, Pandemic and Points of Concern

Time-series remote sensing inventory of Sundarbans mangrove damage from cyclones can be updated fairly rapidly post-storm, and relies less on overcoming the destructive impacts to human infrastructure to facilitate expedient decisions during times of compounded disasters. Updates on spatial mangrove recovery and delayed mortality from Titli, Fani, and Amphan are currently underway despite COVID-19. Indeed, the impacts of successive cyclones to mangroves are visibly more effectual in our imagery the longer the gap between storms as recovery is better enabled with longer temporal gaps, e.g., [17]. However, impacts to humans and wildlife do not necessarily follow that same trend.

Another perspective of this research is that scientists map mangrove damage in most cases, but from a sustainability context, the pandemic caused (and continues to cause) many issues that we describe. Timely mapping efforts should be part of disaster response, and not limited to science advancement years after. We are not the only scientists who have noted this, and we want to get this message out when other have not. News headlines rarely focus on lesser-known regions of the world. Quick action is critical to sustainable systems management, and the pandemic is a barrier in need to future accommodation through disaster planning. We may compare our story, e.g., with the following very recent headline, “Tonga races to prevent a ‘tsunami of covid’ as rescue efforts begin after underwater volcano disaster” [45]. This made headlines quickly, but cyclones are all too often not considered in the same light, despite their proportionately higher destructive frequencies and human mortality.

4. Conclusions

Whilst cyclones continue to affect coastlines in developing countries, quick responses facilitated by remote sensing [32] as well as for quantifying resilience/recovery will become increasingly more critical. The decrease in NDVI could just be a short-term effect (leaf loss, structural damage) and forest recovery may occur quickly in some areas. With longer-term data or monitoring effort, these metrics can be used to characterize and map resistance (no change), resilience (recovery to a previous condition), or regime shifts (original ecosystem loss). For example, after Super Cyclone Haiyan, The Philippines’ Department for Environment and Natural Resources proposed a $22 million (USD) mangrove replanting program for the affected mangrove-fringed coastal zone to develop the first line of defense against future cyclones [13]. Conservation and protection of mangroves alone cannot fulfill future situations for local coastal communities as well as achieving sustainable development goals. The livelihood is definitely an important part of every country. For example, Bakhawan Ecopark, a mangrove reforestation project, did not only address the community’s flood problems but also provided a means of livelihood for the local people in the area through eco-tourism, and has been evaluated as one of the good forest management practice in the whole of Asia and the Pacific by FAO [46]. Thus, mangroves act not only as bio-shield but also provide other functions including carbon sequestration, enhancement of fish stocks, eco-tourism spots, etc. Adding coastal blue carbon ecosystems in Nationally Determined Contributions (NDCs) would be one of the ways for coastal countries to achieve minimization of carbon emissions via protecting mangroves, and create additional opportunity for economic recovery that consider the reality of successive cyclone strikes. Finding ways to maintain mangrove ecosystem health and resilience via natural recovery, reforestation, conservation, and proper sustainable management after successive cyclones must be a part of national-scale strategies in spite of pandemics and future unforeseen circumstances. Necessary delays from pandemics do not change natural resource needs.