Hurricane

The Royal Netherlands Meteorological Institute (KNMI) has been responsible for weather forecasting in the Dutch Caribbean (Bonaire, St. Eustatius and Saba – the BES islands) since 2016. And while weather patterns in the Caribbean often exhibit homogeneous characteristics, this region is also prone to some of the most violent storms on earth in the form of hurricanes. The most infamous example of this is Hurricane Irma (2017), which passed close to Saba and St. Eustatius but made a direct landfall on and severely impacted several other Caribbean islands, including Sint Maarten. Over 90% of the buildings on St. Maarten were damaged, including most of the infrastructure on the island. The estimated damage totaled to be around 2.7 billion USD (approximately 200% of the country's GDP).  

With its extensive weather forecasting expertise as a solid foundation, KNMI is now moving towards impact-based forecasting through the development of the Early Warning Centre (EWC). For the BES islands, this means that we will design a hurricane impact model, combining KNMI's forecasting experience with impact modeling expertise nested within academia. With respect to the latter, we follow the traditional risk modeling approach and set up a hazard – exposure – vulnerability type of model chain. In such model chain, it is predominantly the choice of hazard data that determines the nature and applicability of the output data. For instance, (ensemble) forecast tracks provide insights into possible impacts of an imminent hurricane. Similarly, using synthetic hurricane tracks from a statistical model like STORM will result in a full spectrum of risk and associated probabilities. We will also incorporate local knowledge to develop and improve exposure and vulnerability input data. 

In this presentation, we discuss the different input datasets needed to build an impact model, and how the different output products can assist weather forecasters in better understanding the impact of imminent hurricanes in the Dutch Caribbean.

Caribbean flora and fauna have always coped with the destructive forces of hurricanes. However, climate change leading to an increase in their frequency and strength, and because many species have declined in abundance due to anthropogenic causes, a better understanding of how hurricanes effect local populations is essential.

The Quill before and after Hurricane Irma. Photo credit: Hannah Madden

2017 Hurricane Season

The 2017 Caribbean hurricane season was the most intense recorded to date. Both Irma and Maria, category-5 hurricanes, closely passed Sint Eustatius and caused major destruction on the island; reported in this Nature Today article. Although immediately after the storms it was clear that trees were heavily affected and mostly defoliated, understanding which species were affected and to what extent requires time for data collection and comparison. Since 2017, several studies have provided pieces of information in order to understand how local populations coped, or not, with the 2017 hurricane season.

Since 2017, researchers found that >90% of all trees were defoliated by more than ¾, and that especially trees at higher elevations (such as on the Quill volcano) were affected more severely. Another study that focused on the Bridled Quail-dove (Geotrygon mystacea), demonstrated that the population declined by 77% in 2019 compared to pre-hurricane levels. A follow-up study in 2021 (not yet published) recorded a further decline to just 125 individuals, and the Bridled Quail-dove will likely be re-assessed by the IUCN.

Reptiles

Focusing on reptile species, a novel study further aids our understanding of the ecosystem-wide impact that the 2017-hurricane season had on Statian biodiversity:

The new study, focusing on the Critically Endangered Lesser Antillean iguana (Iguana delicatissima), shows that its population decreased by at least 20% during 2017. Comparing sighting and survey data from 2017–2018, the authors found a decrease in both the occupancy and population size of the iguana species. Importantly, no recovery was observed in 2019, suggesting that this already small population needs multiple consecutive years without major hurricanes to recover. Interestingly, similar to Statia’s forests, iguanas at higher elevations were found to have been affected more severely.

Letter Antillean Iguana. Photo credit: Philippa King

Importance

Small islands such as Sint Eustatius are home to declining populations of rare and endangered species. In many cases, these isolated populations are unable to migrate between islands and thus populations can only increase in size locally. These new studies highlight the need to improve habitat quality and lower anthropogenic threats to optimize the natural recovery of these species. Ideally, at least for population increase could be aided by a local head-starting project where baby iguanas are nourished in temporary captivity and released once they are larger and more likely to survive.

You can find the full study here entitled “Hurricane-induced population decrease in a Critically Endangered long-lived reptile” using the DCBD link below.

More info in the Dutch Caribbean Biodiversity Database
 

Downloads & links >

Published in BioNews 54

Abstract

Post-disaster building damage assessment is an important application of remote sensing. The increasing resolution of remote sensing imaging systems and state-of-the-art deep learning networks has facilitated damage assessment. However, most existing methods in the literature concentrate on damage/non-damage classification only in specific disaster types/areas using pre- and post-event images. Furthermore, site visits are inevitable to assess the level of damage to structures. Therefore, the main objective of this study was to utilize deep transfer learning over a pre-trained network and extend it to a damage assessment framework. The network is fine-tuned to identify four different damage levels: non-damage, minor damage, major damage, and collapsed, using only post-event images taken from different disaster types/areas. To evaluate the proposed framework, we carried out three experiments on Hurricane Irma in Sint Maarten, Hurricane Dorian in Abaco Islands, and Woolsey Fire using post-event orthophotos derived from unmanned aerial vehicle (UAV) images. The results of over 80% overall accuracy confirm that with a structured learning scenario, it is possible to use transfer learning on very high-resolution remote sensing images to classify the level of structural damage.

Full document can be requested here: https://cdnsciencepub.com/doi/10.1139/geomat-2021-0014

We estimated occupancy, abundance (lambda), detection probability, density/ha and abundance of a regionally endemic snake in the Colubrid family on the Dutch Caribbean island of Saba in 2021, four years after hurricanes Irma and Maria impacted the island. Line transect surveys were conducted at 74 sites covering 6.7 ha. The proportion of sites occupied was estimated at 0.74 (min 0.48, max 0.90), with occupancy varying between vegetation types and across elevational gradients. Similarly, lambda was estimated at 1.61 (min 0.7, max 3.7) but varied between vegetation types and elevational gradients. Detection probability was estimated at 0.15 (min 0.10, max 0.21). Using Distance sampling, we estimated 10.9 (min 7.3, max 16.2) racers/ha, with a total population estimate of 4,917 (min 2,577, max 6,362) across the entire study region (438.6 ha.) Based on anecdotal observations from Saban residents and prior literature describing the pre-hurricane population as “abundant” (at least 2.0 racers/hour), we posit that the population experienced a hurricane-induced decline but may have since recovered, though not to previous levels (1.28 racers/hour). Nevertheless, our results suggest that racer densities on Saba are currently higher than those on St. Eustatius. Despite this, given the species’ extremely limited extant range and the presence of invasive species on both islands, prevention of local extirpation should be a high conservation priority.

ABSTRACT

Catastrophic events, like hurricanes, bring lethal conditions that can have population-altering effects. The threatened Caribbean dry forest occurs in a region known for its high-intensity hurricane seasons and high species endemism, highlighting the necessity to better understand hurricane impacts as fragmentation and clearing of natural habitat continues. However, such studies remain rare, and for reptiles are mostly restricted to Anolis. Here we used single-season occupancy modeling to infer the impact of the intense 2017 Atlantic hurricane season on the critically endangered Lesser Antillean Iguana, Iguana delicatissima. We surveyed 30 transects across eight habitats on St. Eustatius during 2017-2019, which resulted in 344 individual surveys and 98 iguana observations. Analyses of abundance and site occupancy indicated both measures for 2018 and 2019 were strongly reduced compared to the pre-hurricane 2017 state. Iguanas at higher elevations were affected more profoundly, likely due to higher wind speeds, tree damage and extensive defoliation. Overall, our results indicate a decrease in population estimates (23.3-26.5%) and abundance (22-23.8%) for 2018 and 2019, and a 75% reduction in the number of opportunistic sightings of tagged iguanas between 2017-2018. As only small and isolated I. delicatissima populations remain, our study further demonstrates their vulnerability to stochastic events. Considering the frequency and intensity of hurricanes are projected to increase, our results stress the urgent need for population-increasing conservation actions in order to secure the long-term survival of I. delicatissima throughout its range.

Abstract

We estimated population densities of the red-bellied racer (Alsophis rufiventris) on the Caribbean island of St. Eustatius in 2011, 2018 and 2019 to determine the likely influence of hurricanes Irma and Maria (September 2017), in addition to evaluating abiotic parameters which may be correlated with its presence. Surveys were conducted at seven sites in 2011 prior to the hurricanes, and at 81 and 108 sites in 2018 and 2019 respectively posterior to the hurricanes. A total of 8.2 ha was surveyed in 2011, and 11.42 ha in 2018/2019. The pre-hurricane (2011) racer density estimate was 9.2/ha (min 7.3 - max 11.6); post-hurricane estimates were 4.6/ha (min 3.4 - max 6.0) in 2018 and 5.0/ha (min 3.8 - max 6.5) in 2019. The pre-hurricane encounter rate of individual racers was 16.0 snakes/hour compared to 0.34 snakes/hour in 2018 and 0.41 snakes/hour in 2019 (post-hurricane). The decrease in encounter rates between 2011 and 2019 implies a negative impact of the hurricanes on racer abundance. Based on calculations of detection probability (0.02 in 2018 and 0.03 in 2019), post-hurricane lambda estimates were 1.82 (95% CI 0.66 - 5.01) in 2018 and 1.60 (95% CI 0.39 - 6.65) snakes/ha in 2019. Given the current small size of the remaining population and the presence of invasive species across the snake’s range, this species could be at risk of local extirpation. We suggest conservation actions such as invasive species management and habitat restoration to enable further recovery.

Msc. thesis

Het eiland Bonaire, onderdeel van de benedenwindse eilanden van de Nederlandse Antillen, Hgt op de rand van de Caribbische hurricanegordel. Het risico bestaat dan ook dat het eiland getroffen wordt door een hurricane. Aangezien de westkust van het eiland weinig hoger dan het waterniveau ligt, is er in zo'n geval weinig bescherming tegen de doordringing van golven op de kust. Om de mate van golfaanval en inundatie te bepalen zijn in diverse voorstudies de waterstandsverhoging en golfhoogte voor de kust van Bonaire bepaald met computermodellen. Hieruit is gebleken dat er een significante golfhoogte van ongeveer 3 m te verwachten is met een maximale waterstandsverhoging van 0.5 m. Voor de overgang van de situatie voor het breken van de golf naar de brekerzone en golfoverslag zijn nog experimenten gedaan, maar de verdere doordringing op het land zijn echter geen computermodellen of theorieen voor handen. Om toch deze situatie te kunnen analyseren is er gekozen voor het gebruik van een schaalmodel. Allereerst is een tweetal prototype situaties gedefinieerd die als representatief kunnen worden gezien voor de westkust van Bonaire. Deze situaties verschillen op het punt dat er een situatie met strandmuur en een situatie zonder strandmuur is. Deze prototype situaties zijn vervolgens op schaal1 :15 nagebouwd in een golfgoot, waarbij ook de golfhoogten en waterstanden zijn omgeschaald. Er is vervolgens een meetprogramma samengesteld waarbij variaties zijn aangebracht in de waterstand, de golfsteilheid van de inkomende golven en de ruwheid van het achterland. Na verwerking van de meetgegevens zijn deze met behulp van spectraalanalyse en dimensie-analyse geanalyseerd om de relaties te vinden tussen de in het model gevarieerde parameters en de mate van golfdoordringing en setup op het land. Uit deze analyses zijn vooral de relaties tussen de waterstand en de golfdoordringing of setup naar voren gekomen Vervolgens zijn de gegevens uit de modelsituatie omgeschaald naar de prototype situatie en is er gekeken naar de impactbelasting die de golven kunnen veroorzaken tegen de gebouwen aan de kust teneinde een zone te kunnen bepalen waarbuiten geen schade op zal treden aan deze gebouwen. De golfhoogten op het land bleken maximaal 0.8 m in de eerste meters te worden voor de situatie waar een strandmuur aanwezig is, bij een waterstand van ongeveer 0.5 m, die over een lange afstand landinwaarts aanwezig blijft. Aan de hand van controle berekeningen voor betonconstructies is gebleken dat het risico van schade aan de wanden van de gebouwen niet zo zeer aanwezig is. Eventuele semi-permanente bebouwing als eettentjes of terrasoverkappingen zouden in gevaar kunnen komen, maar zijn indien daarop ontworpen ook te verwijderen voor een hurricane. Ramen en deuren op de begane grond behoeven wellicht extra bescherming. Door de gebouwen boven het niveau van de te verwachten waterstand op het land te bouwen neemt de kans op schade door overstroming ook behoorlijk af.

THE BOTTOM – About 60 percent of Saba’s nature in general was devastated as a result of the recent hurricane Irma. According to Kai Wulf, Park Manager at the Saba Conservation Foundation, the top priorities of the foundation and it’s team of volunteers are clearing the hiking trails and combating further damage by invasive to the land and sea. 

“Of course it looks really really bad right now, but we are looking to the future very positive. Nature has over the centuries learned to cope with such events, so sooner or later the nature will come back. Sometimes these events also act as a renewal,” said Wulf of the state of nature on the island.