Environmental risk assessment of petrogenic hydrocarbon spills in mangrove ecosystems: the Tumaco case study as a baseline, Colombian Pacific

Petrogenic hydrocarbon spills (PHS) are harmful to mangrove ecosystems along tropical coastlines in the short and long term. The aim of this study was to assess the environmental risk of recurrent PHS on mangrove ecosystems in Tumaco municipality, Colombian Pacific. Mangrove characteristics and management aspects led to subdividing the study area into 11 units-of-analysis (UAs) for which threats, vulnerability, potential impacts, and risks were assessed based on environmental factors and the formulation and use of indicators in a rating scale with five categories, which are very low, low, moderate, high, and very high. The results showed that all UAs are highly (64%; 15,525 ha) or moderately (36%; 4,464 ha) threatened by PHS, highly (45%; 13,478 ha) or moderately (55%; 6,511 ha) vulnerable to this kind of pollution, and susceptible to high (73%; 17,075 ha) or moderate (27%; 2,914 ha) potential impacts. The environmental risk was high in 73% (17,075 ha) of the UAs, indicating likely irreversible damage to mangrove ecosystems by PHS, thus pointing to the need of urgent intervention by responsible authorities to ease their recovery and conservation. The methodology and results of this study become technical inputs that serve for environmental control and monitoring, which can be incorporated into contingency and risk management plans.


Introduction
Pollution of coastal and marine ecosystems is a primary global concern and, as such, was considered in the Sustainable Development Goal 14: Conserve and sustainably use the oceans, seas, and marine resources for sustainable development, which included targets for significantly reducing pollution of all kinds and protect marine biodiversity on which more than 3 billion people rely for their livelihood (United Nations, 2015).
Petrogenic hydrocarbon spills (PHS) are one of the most harmful types of marine pollution due to high recurrence, the environmental damage caused and socioeconomic implications Duke, 2016). Petrogenic hydrocarbons are compounds associated with products or sources of petroleum; in addition to this type of hydrocarbon, there are also biogenic ones (produced by metabolic processes of organisms) and pyrogenic ones (derived from the incomplete combustion of organic material) (Neff et al., 2005;Wang et al., 2014).
Environmental impact of PHS in coastal areas hosting sensitive ecosystems is often severe . Such an impact depends on a few variables, including the volume spelled, the hydrocarbon type and state of weathering, its dispersal and the extent of the area affected, and the characteristics of the exposed natural elements and substrates (Singh et al., 2020).
Mangroves are coastal ecosystems that develop in intertidal areas in tropical and subtropical environments, where the soil is flooded, unstable and anoxic (Nagelkerken et al., 2008). This ecosystem, which is one of the most productive in the world, provides highly valuable environmental services to coastal human communities, such as the provision of natural products; coastal protection from erosion, floods and extreme weather events; carbon capture and storage; water purification; and also as habitat, spawning and nursery areas for resident and migratory species of ecological and economic relevance, in addition of constituting a resource for ecotourism and recreation (Das, 2020;Lee et al., 2014;Nagelkerken et al., 2008).
Mangroves are sensitive to the effects of PHS. It has been estimated that approximately 1.94 million hectares of mangroves have been affected by PHS globally, with at least 126,000 ha lost (Duke, 2016). Petroleum spills in mangroves cause short-and long-term damage to biodiversity, socioeconomic systems, and human health (de Oliveira et al., 2021;Duke, 2016).
According to da Silva et al. (1997), the changes that a mangrove undergoes after a petroleum spill can be grouped into four stages: (1) an initial impact during ~ 1 year, a period in which the most vulnerable organisms die, such as young propagules and seedlings; (2) short-term structural damage to the ecosystem after ~ 2.5 years involving the death of trees; (3) a mid-term stabilization period after ~ 5 years, a time in which deterioration is minimized, though without signs of recovery; and (4) a long-term recovery, of variable duration, in which the ecosystem progressively improves thanks to recolonization and increasing density of plants and other organisms. Studies show that a mangrove forest where trees die from spilled petroleum takes more than 25 years to recover since this involves recruiting new trees that can take 25 to 30 years to reach maturity (Duke, 2016). However, the complete recovery of the mangrove forest cover may take 55 years or more (Connolly et al., 2020).
Mangroves grow on soils with a high content of organic matter and mud, both easing the retention of petrogenic hydrocarbons, which subsequently poison aquatic organisms (Connolly et al., 2020;Lewis et al., 2011;Liu et al., 2018;Naidoo & Naidoo, 2016, 2017 and ultimately lead to the deterioration of the habitat quality (Duke, 2016;Garcés-Ordóñez & Espinosa, 2019). This endangers food security and reduces natural elements that are essential for the livelihood of local human communities, may substantially decrease the generation of economic income from fishing, and can trigger social conflicts (Andrews et al., 2021;de Oliveira et al., 2021).
Since 1970, PHS have shown a global decreasing trend parallel to the improvement of the measures to deal with these contingencies . This problem continues to be recurring in Colombia because of the persistent armed conflict, with more than 2,500 intentional blasts, from 1980 to 2018, to the pipelines carrying oil from inland to the country's coastline. This has caused the spill of ~ 3.7 million barrels of petroleum (Ecopetrol, 2015;Guerrero-Useda, 2018Palao et al., 2021) and led to the pollution of water resources and associated terrestrial, coastal, and marine ecosystems (Angulo-Cuero et al., 2021;Ecopetrol, 2015;Garcés-Ordóñez & Espinosa, 2019).
The establishment of measures and the implementation of actions to prevent or mitigate the impacts of such pollution require sound background knowledge of the level of risk to which natural elements are exposed in terms of the occurrence probability of given adverse effects on ecosystems and people (EEA, 2022;Zhukinskii, 2003). Several risk assessment methodologies have been developed to do so (e.g., Bock et al., 2018;Fahd et al., 2021;Romo-Curiel et al., 2022;Tena-Chollet et al., 2013, to cite some). These methodologies take into account several variables, such as the particular physical and biological characteristics of the exposed natural elements, the level of threat (i.e. the factor external to the ecosystem that represents a risk), the vulnerability (i.e. the socioeconomic, environmental and/or institutional susceptibility or fragility eventually leading to losses due to a threatening event) and the impacts (i.e. the changes induced in natural and human systems due to a given event or activity, or to a set of those) (EEA, 2022;Gallopín, 2006;Zhukinskii, 2003). However, a methodology focusing on interactions within the entire socialecological system has not yet been developed to assess this environmental risk in mangroves. This represents a limitation in determining PHS's effects and establish appropriate management measures for the ecosystem.
The aim of the current study is to assess the environmental risks posed by PHS on the mangrove ecosystem of Tumaco municipality, Colombian Pacific. To achieve that aim, a methodological model has been developed that includes formulating pressure, state, and response indicators and analyzing threats, vulnerability, and potential impacts of PHS on the ecosystem. Hopefully, the results of this study will contribute to the establishment of management measures to prevent or mitigate pollution due to PHS, thus helping the conservation of mangroves. Finally, we expect this study to serve as a basis for similar assessments of the status of mangrove ecosystems elsewhere, given their protection.
Most of the ~ 257,000 inhabitants of Tumaco live in extreme poverty with little access to education, health, and social benefits, and are largely dependent on resources provided by the mangrove ecosystem (DANE, 2018;Alcaldía de Tumaco, 2019). Most housing in rural areas consists of basic stilt houses ("palafitos") on the banks of rivers, and human mobility and transport of goods are made with small boats with outboard motors (Alcaldía de Tumaco, 2019). The main economic activities are fishing, tourism, agriculture, mining, and the operation of the Tumaco seaport, from where the petroleum that arrives through the Trans-Andean oil pipeline is exported (Garcés-Ordóñez & Obando, 2019).
According to the classification of Köppen, the climate in Tumaco is tropical rainforest, with multiannual average temperature from 24 to 28 °C, rainfall from 2,500 to 5,000 mm, and relative humidity from 85 to 90% (IDEAM, 2014a(IDEAM, , 2014b. The tidal regime is semidiurnal, with a maximum height of 3.7 m in springtides (IDEAM, 2010). Coastal currents are between 0.04 and 0.3 m s -1 in speed with a predominant south-north direction from June to August and a west-south direction from February to April (Espinosa, 2017;UNINORTE, 2013).
Twelve main rivers open to the sea along the 400 km of coastline of Tumaco, among which Mira River stands out for its flow that ranges from 800 to 2,000 m 3 s -1 (Espinosa, 2017;INVEMAR, 2022). It is estimated that 12% of the coastline suffers from erosion while 7% is accreting in response to sediment supply and hydrodynamics, the later also influencing substrate characteristics and the retention and dispersal of sedimentary particles and pollutants such as petrogenic hydrocarbons (Espinosa, 2017;Posada et al., 2009).
The most representative coastal ecosystem in Tumaco is the mangrove, which is dominated by tree species of the Rhizophora genus. The most common species are Rhizophora mangle, R. harrisonii, Avicennia germinans, Laguncularia racemosa, Pelliciera rhizophorae, Mora oleifera and Conocarpus erectus (Tavera, 2014). The soil of the mangroves is organic matter rich (i.e., 107 ± 6 mg C g -1 dry weight) with predominating clay (96.5%) and silt (57.7%) grain size fractions (Garcés-Ordóñez & Espinosa, 2019). 26 crustacean and 11 mollusk species of commercial importance have been reported in the mangroves of the study area (Espinosa, 2017). Among the mollusks are "piangüas" A. tuberculosa and A. similis that inhabit the roots of Rhizophora spp., which are a traditional food for local families and are sold in the local, Ecuadorian, and Peruvian markets (Cano-Otalvaro et al., 2012;Gil-Agudelo, 2010).
The mangroves of Tumaco have a management plan in which general guidelines are defined altogether with areas of sustainable use, preservation, and recovery (Tavera, 2014). Sustainable use areas, where wood collection and fishing are allowed, occupy ~ 50% of the total cover of mangroves. Preservation areas, where ecotourism is admitted, cover ~ 35% of the mangrove cover. Finally, recovery areas extend over ~ 15% of the mangrove cover and correspond to degraded areas where the artisanal collection of mollusks is allowed but the commercial extraction of mangroves is prohibited (INVEMAR, 2014;Tavera, 2014).
In November 2017, the Cabo Manglares Bajo Mira and Frontera Integrated Management National District was declared, under the administration of the National System of Protected Areas (SINAP; Fig. 1). This protected marine and coastal area is located south of Tumaco municipality, covering a total area of 190,282 ha. The area encompasses mangrove (2.6%; 4,894 ha) and continental (5.3%; 10,088 ha) ecosystems, and a large marine Sect. (92.1%; 175,300 ha) (PNNC, 2023).

Methodology
The assessment of the environmental risk was carried out in two stages. The first involves the environmental characterization of the mangrove ecosystem by establishing a set of indicators considering pressures, state, and responses, which were subsequently integrated into the threat and vulnerability analysis allowing identifying potential impacts. The second stage consisted of analyzing and attributing values to threats, vulnerability, potential impacts, and environmental risks (Fig. 2).
For the environmental characterization and formulation of the indicators, we reviewed all technical and scientific available literature with information about socioeconomic activities, coastal dynamics, petroleum hydrocarbon concentrations in water, sediments, and aquatic organisms, and flora and fauna species. Documents were searched in bibliometric databases such as Scopus and Google Scholar and requested from local and national public and private institutions. Also, the study area's biophysical and chemical measurements were analyzed (Espinosa, 2017).
The information assessed led to the formulation of 36 indicators of anthropogenic pressures (i.e., extraction of resources and pollution of the ecosystem), status (i.e., quality, quantity, and sustainability of resources) and response/management. These simple indicators of level I were grouped into six criteria (partial index of level II) and nineteen attributes (partial index of level III) for threat and vulnerability analyses (partial index of level IV; Table S1A-B). A rating scale with a set of ranges was applied to these simple indicators of level I that, under ecological risk assessment (EEA, 2008), combines quantitative and qualitative data about the environmental variables. The criteria for the analysis of potential impacts (partial index of level IV) and environmental risk (synthetic index of level V) were also defined (see subsections 2.2.3 and 2.2.4; Table S1).
To integrate the values of the simple indicators of level I, attributes, and criteria defined for the analysis of threats, vulnerability, potential impacts, and environmental risk, the weighted geometric mean formulas (Eq. 1) were used, which allows expressing their established relative importance values (weightings), where each of them intervenes differentially in the result. This equation is considered the best and most representative aggregation function widely used in formulating environmental indices (Khouri & Al-Moufti, 2022;Singh et al., 2008;Zardari et al., 2015). With it, the aggregation of the simple indicators of level I values is done, passing to the partial indices of level II (attributes), III (criteria), IV (components: threats, vulnerability, and potential impacts) and the synthetic index of level V (environmental risk; Table S1). The simple indicators of level I values correspond to the rating scale defined in Table S1. where: • j = indicators, attributes, criteria, or components that make up the index and that take values between 1 and m • m = number of indicators, attributes, criteria, or components • nj = value or score of the indicator, attribute, criterion, or component analyzed j • wj = weighting factor of indicator, attribute, criterion, or component j The attributes (partial index of level II) for each criterion by UA were calculated with Eq. (2): where: • At j = attribute (partial index of level II) calculated separately for each UA • n j = rating scale of indicator j analyzed • w j = relative weighting factor of each indicator (Table S1).
Once the attributes were estimated (partial indices of level II), the calculation of the criteria (partial indices of level III) continued, using Eq. 3: where: • Cr j = criteria (partial indices of level III) calculated separately for each UA • n j = result obtained from the attribute (partial index of level II) • w j = relative weighting factor of each attribute (Table S1) After having estimated the criteria (partial indices of level III), the calculation of the components (partial indices of level IV) proceeds, for which Eq. 4 was used: where: • Coj = components (partial indices of level IV) calculated separately for each AU • nj = result obtained for each criterion (partial index of level III) • w j = relative weighting factor of each criterion (Table S1).
The weightings of the simple indicators, attributes, criteria, and components in the formulas used were established by the expert judgment of the research team and other experts who were also consulted. Finally, a qualitative rating scale of five categories was established -very low, low, moderate, high, and very high-, together with their respective interpretations (Table 1).

Threat component analysis
The simple indicators, attributes, and criteria defined for calculating the threat index, and their respective weightings and rating scales, are shown in Table S1A. Threat component levels were calculated using the Eq. 4 using the criteria of probability (i.e., the level of certainty that a PHS will occur), magnitude (i.e., the intensity of the PHS), and extension (i.e., the spatial or temporal amplitude of the PHS), with weightings of 35%, 30%, and 35%, respectively.
Probability has a high rating because the recurrence of breakages in the Trans-Andean oil pipeline, illegal extractions, and petroleum leaks generate constant spills in the study area (Ecopetrol, 2015;Tejada & Afanador, 2003). The magnitude and extension criteria rating were calculated after the Eq. 3, considering the respective weightings of the defined attributes and indicators (Table S1A). The results of the threat calculation were scored according to the rating scale described in Table 1A. (4)

Vulnerability component analysis
The simple indicators, attributes, and criteria defined for the calculation of vulnerability, and their respective weightings and rating scales, are shown in Table S1B. The vulnerability of the mangrove ecosystem was calculated using the Eq. 4 after the exposure (i.e., the duration of hydrocarbon pollution in the ecosystem), ecosystem sensitivity (i.e., the potential to affect organisms) and the adaptive capacity of the ecosystem (i.e., the capacity of recovery and resistance to hydrocarbon pollution) criteria, with weightings of 40%, 40%, and 20%, respectively. The estimation of each criterion and attribute was also calculated with Eq. 3, using the indicators defined for the vulnerability analysis and their respective weightings (Table S1B). The results of the calculating the vulnerability of the mangrove ecosystem refer to the rating scale shown in Table 1B.

Analysis of potential environmental impacts
Potential environmental impacts are determined considering a scenario where the coastal area of Tumaco is affected by the spill of ∼14,000 barrels of crude petroleum, as it occurred in 2015. The assessment of the potential impacts of PHS on the mangrove ecosystem has been performed by integrating the results of a network method generalized for the entire study area after Martínez-Bernal (2013), and a qualitative method for each UA in the study area following Sikdar (2021).
The network method for impact assessment is based on the direct causal relationships that arise between the impacting actions and the resulting potential impacts. To implement it, the activities with the potential to generate impact were first identified and classified into three groups that are referred to as A1 from poorly visible spills and retail selling of petroleum hydrocarbons, A2 from the operation of the oil terminal monobuoy, and A3 from inland PHS of the Trans-Andean oil pipeline.
A1 refers to small PHS that constantly occur from small boats that transit within the Tumaco cove and its surroundings, which are difficult to quantify and locate (Tejada & Afanador, 2003). A2 refers to the Ecopetrol Oil Terminal, located 8 km from the coastline of Tumaco's urban center. A3 corresponds to spills in the Trans-Andean oil pipeline due to improper manipulation of the shut-off valves, militia attacks or natural events that affect the oil pipelines (Tejada & Afanador, 2003). Subsequently, 29 environmental factors grouped into eight "compartments" were identified: soil, water, atmosphere, flora, fauna, recreation, aesthetics and livelihoods, which could be potentially impacted by the three above-defined groups of activities A1, A2 and A3, together with the potential impacts of these activities on each environmental factor using a double-entry matrix (Table S2; Martínez-Bernal, 2013). Of these "compartments" the first five are purely environmental, "recreation" is a usage with environmental implications, "aesthetics" refers to the human perception that is also linked to environmental aspects, and "livelihood" refers to resource exploitation and, more widely, to ecosystem usage, also with environmental derivatives.
Then, the causal relationship was analyzed using an adjacency matrix with the activities A1, A2 and A3 and the potential impacts identified, using 1 or 0 depending on whether or not there is a direct causal relationship between the elements analyzed (Table S3; Martínez-Bernal, 2013). To finalize the network method and determine the potential impact significance, an analysis of the relationships between impacting activities and potential impacts was performed, determining the total number (total degree) of interactions (causes and consequences) using the Netdraw function of the UCINET software, version 6.631, from which the levels of significance established by Martínez-Bernal (2013) were qualified in the categories of critical, severe, moderate, and irrelevant.
The qualitative method was used to prioritize the mangrove ecosystem's environmental components, presenting the greatest potential impacts by UAs in the study area (Sikdar, 2021). Considering the expert judgment, a weighting was assigned to each of the above-mentioned eight environmental compartments and category of impact significance level in the network method, and it was reclassified in an impact rating scale from 1 (very low), 2 (low), 3 (moderate), 4 (high), and 5 (very high). To finalize such a qualitative approach, Eq. 2 was used to integrate the assessments of each environmental compartment to mangrove UAs. The integration of results on potential impacts from the network and qualitative assessments was made by weighting the respective outputs in Eq. 5. Finally, the outcomes obtained were classified according to the rating scale shown in Table 1C. where: • Pi j = results of potential impacts calculated separately for each AU • nj = result obtained with the network method (Nm) and qualitative method (Qm) • w j = relative weighting factor of each method (Nm = 0.30; Qm = 0.70)

Environmental risk index analysis
The synthetic level V index of environmental risk due to the PHS was determined after the relationship between the probability of occurrence of the threat, the vulnerability of the affected UA in the mangrove ecosystem, and the susceptibility of environmental factors to the potential impacts derived from PHS in the coastal area of Tumaco (Tejada & Afanador, 2003;Toro et al., 2012), according to Eq. 6. The risk assessment results were classified into five categories according to Table 1D where: • ERj = environmental risk calculated separately for each UA • nj = result obtained from the partial indices of level IV (components of threat, vulnerability, and potential impacts) • w j = relative weighting factor of each component (Table S1) Map production To produce the maps with the results of the threat, vulnerability, potential impact and environmental risk analysis, baseline information was collected for the study area (INVEMAR & IPC, 2013;Tavera, 2014;INVEMAR, 2016; IGAC, 2019) that allowed to define the attributes, indicators, and rating scales for the analyses. The geographic information collected passed through a reviewing and reliability assessment process, and vector layers were selected considering temporality, scale, and source. An updating process was applied to the information on mangrove coverage at a scale of 1:100,000 using the 2015 Rapideye satellite image, geometrically corrected. For this, the correspondence of image pixels with reality was evaluated by cross-checking with GPS points and route lines measured in the field. Also, elements defined by Weng (2010), such as tone, size, shape, texture, and pattern were considered to identify plant cover. The needed polygon adjustments were made manually by using the Arcgis 10.3 software. For these adjustments, we used the combination of false color bands R = Near Infrared, G = Limit Red, and B = Red. Then, the geographic information went through a topological review process to avoid gaps and overlaps between polygons that could have resulted in erroneous data in terms of extension, thus ensuring the quality of the geographic information in the maps.
The 11 spatial UAs (Fig. 1) were defined from the updated layer of mangrove extension at 1:100,000 scale aided by expert judgement, thus allowing the application of the different methods of analysis using the established indicators. These analyses were carried out with geoprocessing tools permitting the qualification matrices to be integrated into the database.

Threats component
The rating results of the indicators and criteria for threat analysis of the mangrove ecosystem are presented in Table S4 and Table 2, respectively. Probability had a very high rating for 100% of the UAs; magnitude has a low rating in 55% of the UAs, moderate in 36% and high in 9%; and extension is moderate and high in 36% and 64% of the UAs, respectively. The integration of results for these criteria shows that 64% (15,525 ha) of the UAs (Curay, Trujillo, Coba, Tumaco, Bocagrande, Terán and Congal) have a high threat of petroleum hydrocarbon pollution. In contrast, the other UAs (36%; 4,464 ha) have a moderate threat (Fig. 3).

Vulnerability component
The rating results of the indicators and criteria for vulnerability analysis of the mangrove ecosystem are presented in Table S5 and Table 2, respectively. The exposure to petroleum hydrocarbon pollution is moderate in 64% of UAs, high in 27% and very high in 9%; ecosystem sensitivity is moderate and high in 36% and 64% of the UAs, respectively; and adaptive capacity is low in 18% of the UAs, moderate in 45%, high in 27% and very high in 9%. When integrating these criteria results, 45% (13,478 ha) of the UAs  (Fig. 4).

Potential environmental impacts component
Forty-seven potential impacts by PHS on the eight environmental components of the Tumaco mangrove ecosystem have been determined (Table S2). Of the potential impacts identified and assessed using the network method, 12.5% are critical on soil and 21.4% on water; 12.5% are severe on soil; 12.5% are moderate on soil and 66.7% on flora; and compatible impacts refer to the atmosphere (air quality), fauna, recreation, aesthetics, and livelihoods (Table 3).
The results of the qualitative method show, for all the UAs, a very high potential impact rating for the soil, water, and livelihood mangrove components; high impact rating occurs for flora and fauna components; moderate impacts involve recreational and aesthetic components; and irrelevant impacts refer to the atmosphere component. The network method's integration of the impact rating determined for each environmental component shows a moderate impact on 100% of the UAs. While the integration of the rating of the components determined by the qualitative method shows very high, high, moderate, and low impacts on 64%, 18%, 9% and 9% of the UAs, respectively. Finally, integrating the impact assessments with the two methods applied shows very high impacts on 73% (17,075 ha) of the UAs and moderate impacts on 27% (2,914 ha) of the UAs (Fig. 5).  Table 1 for threat level descriptions Environmental risk index Environmental risk for the mangrove ecosystem in the study area resulting from PHS is moderate in 27% (2,914 ha) and high in 73% (17,075 ha) of the UAs, respectively (Fig. 6). Moderate levels of environmental risk correspond to the Curay, Resurección and Colorado UAs.

Outcomes from threat assessment
The mangrove ecosystem of Tumaco is highly to moderately threatened by PHS depending on the UA considered ( Figs. 1 and 3 to 5). Spills occur as a consequence of a rather large variety of specific events involving va-riable volumes of oil, such as those caused by shipwrecks, operational accidents the oil terminal monobuoy, overflow of storage tanks, and attacks to and illegal extraction from the Trans-Andean oil pipeline, and also by constant dumping of oil waste from retail fuel sellers and boat traffic of boats, among others (Bermúdez & Corredor, 2006;Cabrera & Reyna, 1997;Espinosa, 2017;INVEMAR, 2015). It is to be noted that similar threats have been reported in other mangrove systems all over the world, like the Niger Delta in Nigeria (Iturbe-Espinoza et al., 2022;Paschaline & Sam, 2020), the Spencer Gulf in South Australia (Connolly et al., 2020), Ho Chi Minh City in Vietnam (Khoi et al., 2023), the Gulf of Paria in Trinidad (O'Brien-Delpesh et al., 2022) and the coast of Mauritius in the Indian Ocean (Rajendran et al., 2022). The extension of hydrocarbon pollution in the study area is high to moderate, depending on the UAs considered. It depicts spatial and temporal fluctuations related to spills and the effects of weathering processes on different hydrocarbon types and coastal dynamics (Espinosa, 2017). Tides, currents, waves, winds, and rainfall contribute to the Table 3 Potential impact ratings (in percentage, n = 44) by petrogenic hydrocarbon pollution on mangrove's environmental compartment against impacting actions using the network method. "Recreational" is not an environmental component but a usage.
"Aesthetics" is a human perception, not a usage. "Livelihood" refers to resource exploitation and, more widely, to ecosystem usage, and is not an environmental compartment per se

Compartment
Compatible ( Table 1 for potential impact level descriptions dispersal and boosting of the environmental effects of PHS and to their dilution over larger areas as well (Al-Majed et al., 2012;Espinosa, 2017;Platónov & Redondo, 2003;UNINORTE, 2013). Petroleum spilled both inland and at sea easily reaches the mangroves favored by the high hydrological connectivity of the area's river-estuary-ocean systems and the tides' dynamics (Espinosa, 2017;ITOPF, 2011;Singh et al., 2020). Several rivers, including large ones, flow into Tumaco cove, where in connection to river discharge and metoceanic conditions riverine plumes spread along the coast and seawards (Bastidas et al., 2008). As the tides move up and down, estuarine systems fluctuate between partially mixed and well mixed (UNINORTE, 2013). At the Mira River mouth during tidal ebb, the velocities of the freshwater upper layers from fluvial contributions are larger, which facilitates the movement of hydrocarbons towards the coastal zone, and then their dispersion in a northeast direction (INVEMAR, 2015;Liu et al., 2018;UNINORTE, 2013). During the tidal flow, hydrocarbon dispersion is limited as the salt wedge partly dams the fluvial discharge. The associated increase in the water level causes hydrocarbons to spread towards the mangroves and muddy beaches in them, where part of the spilled oil gets trapped (Espinosa, 2017;UNINORTE, 2013).
The UAs where the magnitude of petroleum hydrocarbon pollution is high and moderate correspond to mangrove areas with organic matter-rich silty and clayey sediments or peat-type soils (Espinosa, 2017;Garcés-Ordóñez & Espinosa, 2019). Small muddy beaches that form on the banks of the mangroves also have high  Table 1 for Environmental risk level descriptions affinity for hydrocarbons (Garcés-Ordóñez & Espinosa, 2019;Olguín et al., 2007).
Pollution by petroleum hydrocarbons in water, sediments and organisms has been reported in the investigated UAs. INVEMAR (2015) noticed dissolved and dispersed hydrocarbon concentrations ranging from 17.4 to 170.7 μg L -1 , with the highest values found in Trujillo, Coba and Bocagrande UAs. Espinosa (2017) found concentrations of aliphatic hydrocarbons of up to 70 μg L -1 in surface waters, with higher values occurring at the mouths of the Mira and Rosario rivers (Fig. 1), which received the petroleum spilled in June 2015. Garcés-Ordóñez and Espinosa (2019) identified high pollution by petroleum hydrocarbons (> 200 µg g -1 in chrysene equivalent) in mangrove sediments of Coba UA and moderate pollution (> 50-200 µg g -1 ) in Bocagrande and Congal UAs. In these very same mangrove UAs, concentrations of aliphatic hydrocarbons in sediments of up to 15.164 μg g -1 dry weight were recorded (Espinosa, 2017).
Other studies carried out in Tumaco have reported dissolved and dispersed hydrocarbons in the waters near fuel service stations and the monobuoy of the oil terminal (Marrugo, 1990;Marrugo & Palacio, 1991;Garcés-Ordóñez & Obando, 2019). ∑16 PAHs have also been found in the sediments of the estuaries of the rivers flowing into Tumaco cove Espinosa, 2017) and in the mangroves of Coba and Bocagrande UAs, which in some cases exceed the reference levels for adverse effects on the aquatic biota (Garcés-Ordóñez & Espinosa, 2019). The presence of chrysene, fluoranthene, phenanthrene and pyrene has been detected in mollusks (Anadara spp.) from the same mangrove UAs with high water and sediment pollution by hydrocarbons (Espinosa, 2017).

Outcomes from vulnerability assessment
The high and moderate vulnerability of the mangrove to petroleum hydrocarbon pollution is indicative of the ecosystem's elevated exposure and sensitivity to this type of pollution, and of its adaptive capacity. Coba, Tumaco, Bocagrande and Congal UAs present the highest exposure values given the dominance of clayey sediments that allow for a high residence time of hydrocarbons. These sediments are characterized by high porosity and reactivity to pollutants' adsorption, which fixed into the soil/sediments in 3 to 10 years (Blumer & Sass, 1972;Botello et al., 2005;Porta et al., 2014). Also, the above UAs are directly influenced by the discharges of the channel networks of Mira and Rosario rivers, which were the main recipients of the oil spills that occurred in June 2015 (Espinosa, 2017;Posada et al., 2009).
The exposure of the mangrove to crude oil in adjacent low-energy beaches increases from days to months since the vegetation along the riverbanks slows down natural removal rates, thus enhancing the ecosystem's vulnerability (Bejarano & Michel, 2016). Instead, in high-energy beaches adjacent to the mangrove, natural removal of oil can be fast, usually in the order of days or weeks (Núñez-Solís, 2013;Olguín et al., 2007;Petersen et al., 2002;Sivagami et al., 2019;Zamora et al., 2012).
Mangrove sensibility in Tumaco is high in 64% of the UAs due to high levels of human intervention and the presence of species that are conservation targets and also are highly sensitive to environmental changes, including those from recurrent PHS, reported here (López-Rodríguez et al., 2008). This is well illustrated by the finding of high levels of petroleum pollution in sediments from different areas within the mangroves of Tumaco two years after the June 2015 spills (Garcés-Ordóñez & Espinosa, 2019). PHS deteriorates the habitat of multiple species that nest, reproduce and feed in the mangrove and causes sublethal effects in some organisms (Duke, 2016;Núñez-Solís, 2013). Such effects vary over time since petroleum hydrocarbon pollution can persist for weeks, months, or years before being naturally removed after getting trapped in organic matter-rich fine-grained sediments (Feito, 2007;Núñez-Solís, 2013).
Areas within mangroves of Tumaco that are zoned for preservation (Fig. 1) have a high adaptive capacity as they keep good structural indices, medium to high productivity, low anthropic intervention, and are far from human settlements (Tavera, 2014), which overall reduces vulnerability. On the other hand, mangroves in sustainable use areas (Fig. 1) experience higher levels of intervention that reduce their adaptive capacity and augments their vulnerability. Other authors also support the view that preservation areas have a shorter recovery time from disturbance events (Jacob et al., 2018;Jones & Schmitz, 2009).
Outcomes from potential environmental impacts assessment Petroleum hydrocarbon spills, depending on their magnitude and extension, can trigger various negative impacts on the mangrove ecosystem and its food web, from the cellular level to given species of fauna and flora, from the ecosystem structure to the quality of environmental matrices (Duke, 2016;Naidoo & Naidoo, 2016, 2017Renegar et al., 2022). These pollutants can be resuspended in the water and transported elsewhere by currents and tides. They can also be oxidized and transformed into more toxic metabolites worsening adverse effects on the mangroves and on human health (Adams et al., 2008;Cai et al., 2019;Garcés-Ordóñez & Espinosa, 2019;Kingston, 2002;Zambrano et al., 2012). PHS potential impacts on the mangrove ecosystem of Tumaco can be high and moderate, resulting in damage that would require mitigation and recovery actions in the short (1-2 years), medium (2-5 years), and long term (> 5 years). The negative impacts of PHS can be cumulative and recurrent over time.
Structural damage by petroleum hydrocarbon pollution at the cellular level may affect roots and leaves, modifying their physiological functions and causing mutagenesis (Naidoo & Naidoo, 2016, 2017. Such pollution may also induce necrosis, perforation, canopy thinning, wilting, leaf yellowing and defoliation (Duke, 2016;García & Martins, 2021;Renegar et al., 2022). When it gets chronic, impacts can result in tree death, reduced growth of surviving trees, seedling degeneration and poor seedling development (García & Martins, 2021;Pavanelli & Loch, 2018;Renegar et al., 2022). In Spencer Gulf, South Australia, oil pollution caused total or severe defoliation of mangrove trees in 3.2 ha after 3.5 years of the spill (Connolly et al., 2020), and in Panama, where two oil spill events oil spills killed approximately 118 ha of this forest (Duke, 1997).
As for phytoplankton and zooplankton, they can die from petroleum pollution, thus transforming otherwise more or less colored and turbid water into crystalline water with a false appearance of cleanliness (Vera et al., 2008). Marine organisms experience different toxic effects, such as disorders in their physiological behavior that could lead to direct mortality, manifestation of sublethal toxic impacts due to carcinogenic effects, mutagenesis, and adverse effects on the development of aquatic species, such as alterations in their reproduction, abundance, and diversity (Duke, 2016;Renegar et al., 2022;Santana et al., 2018;Sivagami et al., 2019;Zhang et al., 2019).
Outcomes from environmental risk assessment Petroleum hydrocarbon spills in Tumaco imply a moderate to high probability of generating severe damage to the mangrove ecosystem in all UAs, since this pollution, especially at high levels, is of persistent character and, therefore, requires the implementation of mitigation measures and restoration plans wherever needed (Espinosa, 2017;Garcés-Ordóñez & Espinosa, 2019;INVEMAR, 2015). A large part of past spills in Tumaco and in other areas of Colombia have occurred within the framework of the internal armed conflict, and although peace agreements between the national government and the guerrillas were signed in 2016, attacks to oil pipelines have continued, thus significantly contributing to persistent environmental damage (Castro et al., 2019;Vélez-Torres & Méndez, 2022).
The damage to the mangrove has repercussions over the production systems, the economy, and the livelihood of the local populations (Ecopetrol, 2015). This is a particularly worrying situation as human communities in the Tumaco municipality largely subsist on the resources of the mangrove and the rivers feeding it (Cámara de Comercio de Tumaco, 2020).
The area's main activity of human settlements is fishing, which has been constantly affected by this type of pollution (Ecopetrol, 2015;Espinosa, 2017). This poses a potential risk to the health of locals, as they can be affected in the short and long term by exposure to and ingestion of hydrocarbon polluted water and food, as reported from other areas affected by PHS and the ensuing pollution (Fadigas, 2017;O'Brien-Delpesh et al., 2022;Paschaline & Sam, 2020).
The methodology used in this study has the advantage of being flexible, allowing it to be adapted to the environmental conditions of the affected area, integrating different ecological, social, and environmental management aspects that allow the identification and zoning of threats, vulnerabilities, environmental impacts, and the risk, for a better response to a disaster event or environmental emergency. Therefore, we can consider it an environmental management instrument for preventing and addressing the risk of PHS.
Also, this methodology can be applied to other mangroves affected by PHS and other types of pollutants, such as those mentioned in Sect. 4.1 and beyond. However, the outcomes could be limited by the lack of environmentally relevant data of the areas to be assessed, so that eventually costly baseline studies could be needed. Information on the levels of ecological risks from pollutants' concentrations in water, sediments, and organisms can be integrated into this methodology to design pressure, state, and response indicators. In particular, the methodology applied, and the results obtained in our study could be of value for the specific case of Cabo Manglares marine protected area, as they provide background information and depict guidelines to the authoritis in charge, specilly in the design of risk management and contingency plans.

Conclusions
The management of environmental authorities in developing countries is limited by little knowledge and lack of management tools that guide the actions required in the most vulnerable areas in the face of threats of marine pollution, such as crude oil spills in tropical coastal areas. This paper assesses the environmental risk of PHS in the mangrove ecosystem of Tumaco. In this Colombian Pacific coastal municipality, the human population suffers from poverty and subsists on mangrove resources and services. Risk has been evaluated after considering the threats posed by recurrent hydrocarbon spills, the vulnerability of the mangrove ecosystem itself, and the potential impacts of this type of pollution.
The vulnerability of the mangrove ecosystem to PHS is high or moderate in areas hosting conservation target species, with elevated sensitivity and levels of human intervention, which make them more exposed to this pollution, therefore reducing their adaptive capacity. Since the environmental impact of PHS on the mangrove is actually or potentially high, short, mid, and long-term measures are required to either maintain or recover the ecosystem face to this type of pollution.
The risk for new petroleum spills to occur in Tumaco is high, which will add to previous impacts in a cumulative way thus further affecting the mangrove ecosystem and the local population. Such a risk is favored by the hydrodynamic processes in the area and the high interconnection of water bodies (rivers-estuaries-sea systems), which ease the dispersion of petroleum hydrocarbons. This study contributes to the identification and spatial lay out of the areas (or UAs) most threatened by oil spills, the threats' sources, and the levels of environmental risk, thus providing much needed background data for the design and implementation of preventive and mitigation measures and contingency plans. Likewise, it contributes with a methodological development that can be applied in other mangrove areas of the world potentially affected by oil spills, to better manage the conservation and protection of these sensitive ecosystems.
Given the persistence of the risk of PHS in Tumaco and its long-lasting effects on ecosystem resources and services to local communities, and in application of the precautionary principle, the above-mentioned measures and contingency plans must deserve priority to ensure a good environmental status of the mangrove ecosystem and a better, healthy future to the local population.