SciELO - Scientific Electronic Library Online

 
vol.50 issue1Monitoring of the status of eleven bony and cartilaginous fish populations in the Malpelo Sanctuary of Fauna and Flora, Colombian PacificGeneral description of some meteo-marine variables of surface waters adjacent to the Palmer Archipelago, Antarctic Peninsula, during 1979-2018 based on Era5 data author indexsubject indexarticles search
Home Pagealphabetic serial listing  

Services on Demand

Journal

Article

Indicators

Related links

  • On index processCited by Google
  • Have no similar articlesSimilars in SciELO
  • On index processSimilars in Google

Share


Boletín de Investigaciones Marinas y Costeras - INVEMAR

Print version ISSN 0122-9761

Bol. Invest. Mar. Cost. vol.50 no.1 Santa Marta Jan./June 2021  Epub Sep 18, 2021

https://doi.org/10.25268/bimc.invemar.2021.50.1.994 

Research Articles

Delimitation and classification of coastal wetlands: Implications for the environmental management of the Colombian Continental Caribbean

Santiago Millán1 
http://orcid.org/0000-0002-4082-937X

Jenny Alexandra Rodríguez-Rodríguez2 
http://orcid.org/0000-0001-8082-8374

Paula Sierra-Correa3 
http://orcid.org/0000-0001-7252-7993

1Instituto de Investigaciones Marinas y Costeras-Invemar, Laboratorio de Servicios de Información (LabSIS), Santa Marta, Colombia. alexandra.rodriguez@invemar.org.co

2Instituto de Investigaciones Marinas y Costeras-Invemar, Laboratorio de Servicios de Información (LabSIS), Santa Marta, Colombia. alexandra.rodriguez@invemar.org.co

3Instituto de Investigaciones Marinas y Costeras-Invemar, Laboratorio de Servicios de Información (LabSIS), Santa Marta, Colombia. paula.sierra@invemar.org.co


ABSTRACT

This article describes the cartographic layer construction process of Colombian Caribbean coastal wetlands at a scale of 1:100,000 and the results obtained in terms of their quantification and typing. Two cartographic layers were constructed and subsequently joined, one of the permanent water bodies and another of temporary water bodies and associated coverages. The layers were generated by multitemporal analysis of 45 Landsat 8-OLI satellite images, based on the NDVI index, uncertainty models by superposition of cartographic attributes, and a flood frequency consultation model on ALOS PALSAR 1 images. As a result, 576,279 ha of coastal wetlands were delimited (1.9 % of total wetlands in Colombia), of which 20.4 % are within protected areas. The cartographic legend makes it possible to typify wetlands based on the coverage and temporality of water bodies; discriminates permanent wetlands (42.7 %) with five categories and temporary wetlands (57.3 %) with 15 categories, mostly distributed in seven large complexes. This study is the first description of the colombian Caribbean coastal wetlands based on a cartographic construction, is methodologically replicable, and will support decision-making in the planning of colombian Caribbean coastal areas, especially for risk management and ecosystem-based adaptation to climate change.

KEY WORDS: Mangroves; Coastal floodplains forest; Water bodies; Geographic Information Systems (GIS); Protected areas

RESUMEN

Este estudio describe el proceso de construcción cartográfica de humedales costeros del Caribe colombiano a escala 1:100 000 y los resultados obtenidos en cuanto a su cuantificación y tipificación. Se construyeron dos capas cartográficas que posteriormente se unieron, una de cuerpos de agua permanentes y otra de cuerpos de agua temporales y sus coberturas asociadas. Las capas fueron generadas mediante análisis multitemporal de 45 imágenes de satélite Landsat 8-OLI, a partir del índice de vegetación-NDVI, modelos de incertidumbre por superposición de atributos cartográficos y la consulta de un modelo de frecuencias de inundación basado en imágenes ALOS PALSAR 1. Como resultado se delimitaron 576 627 ha de humedales costeros (1,9 % del total de humedales de Colombia), de los cuales el 20,4 % se encuentra dentro de áreas protegidas. La leyenda cartográfica obtenida permitió tipificar los humedales con base en la cobertura y temporalidad de los cuerpos de agua; discriminando los permanentes (42,7 %) en cinco categorías, y los temporales (57,3 %) con 15 categorías, la mayor parte distribuidos en siete grandes complejos. Este estudio es la primera descripción de los humedales costeros del Caribe colombiano basada en una construcción cartográfica, es metodológicamente replicable y apoyará la toma de decisiones en la planificación de las zonas costeras del Caribe colombiano, especialmente la gestión del riesgo y la adaptación al cambio climático basada en los ecosistemas.

PALABRAS CLAVE: Manglares; Bosques inundables costeros; Cuerpos de agua; Sistemas de Información Geográfica (SIG); Áreas protegidas.

INTRODUCTION

Inland and coastal wetlands in the world cover more than 12,100 million ha, approximately equivalent to 8 % of the earth’s surface. 54 % of it is permanently flooded and 46 % temporarily; about 92.8 % of them are inland wetlands and 7.2 % are coastal and marine; The South America and Caribbean countries (Neotropical realm) rank third in wetland quantity with 15.8 %, after Asia (31.8 %) and North America (27.1 %) (Davidson et al., 2018); in Colombia, wetlands cover 26 % of the country’s surface corresponding to 30,781,149 ha (Jaramillo et al., 2015).

Coastal wetlands could be permanent or temporary, of saline, brackish or fresh waters, under the direct influence of tides or saltwater intrusions, or atmospheric deposition of substances or particles from the ocean (Ricaurte et al., 2019). They are subjected to changes and degradation by natural and anthropogenic drivers of climatic change, coastal developments, and food production, which cause the most common effects on cultural services, biodiversity, and primary production (Rocha et al. 2015). The increase in the sea level (SLR) can be considered as one of the main factors of climate change that could impact intertidal forests and mangroves (Giri et al., 2011; Alongi, 2015); estimations suggest that a 1m SLE could put into risk 72 % of the world coastal wetlands (Blanckespoor et al., 2014). This could impact ecosystem services that are essential for humanity (Lotze et al., 2006; Worm et al., 2006; Rocha et al., 2015) such as weather regulation through carbon capture and storage, water provision, food and fishing support, different biotic resources, protection against coastal erosion, flood mitigation, recreation and tourism (Liquete et al., 2013; RAMSAR, 2018).

This is why it is necessary to have cartographic supplies that allow supporting the decision-making about ordering and management of coastal zones facing local, regional, or global problems such as climatic change. In the case of Colombia, wetland delimitation was prioritized as an essential planning tool for risk management and adaptation to climatic change after floods during La Niña period in the years 2010 and 2011, encouraging the identification of delimiting criteria based on biotic, geomorphological, pedological, and hydric aspects of the landscape. The cartography of coastal wetlands in the Colombian Caribbean presented here, and built under the directions for the delimitation of Colombia’s continental wetlands included in Vilardy et al. (2014) and Cortés-Duque and Estupiñán-Suárez (2016), is a tool to support the management, planning, and ordering of these ecosystems, as well as to know its representativeness in the National Environmental System (SINA).

STUDY AREA

The Colombian Caribbean coastal line reaches 2,070 km. The coastal zone is located in northwestern South America; to the east, it borders Venezuela at the Castilletes sector (11°50’ N, 71°20’ W) and to the west shares border with Panamá at Cabo Tiburón sector (8°41’ N, 77°21’ W) (Figure 1) (INVEMAR, 2019). The main political and administrative division includes eight departments from south to north as follows: Chocó, Antioquia, Córdoba, Sucre, Bolívar, Atlántico, Magdalena and La Guajira. Regarding the climatic standpoint, in the south end, where the Chocó department is located, it has the most humid weather in the Caribbean, while the aridest one is found in the north end in La Guajira department. The region shows in general, a dry season between December and April marked by the northeast trade winds, a transition season between May and July, and a rainy season between August and November (Andrade and Amaya, 2001).

Figure 1 a Study area framework (coastal area shaded in grey). b. multi temporary visualization through ndvi index. red colors indicate present water in the lower number of dates, blue colors indicate present water in all or almost all dates. c and d pixel values about the uncertainty on wetlands presence. darker tones indicate less uncertainty about wetlands presence. 

MATERIALS AND METHODS

Building a cartographic product for categorizing and delimiting coastal wetlands.

The cartographic building included theoretical conceptualization, satellite image processing, application of geographic information systems (SIG), and on-site information collecting. During the stage of theoretical conceptualization, the starting point was the definition of wetlands by Cortés-Duque and Estupiñán-Suarez (2016) which is: “kind of ecosystem that, due to geomorphological and hydrological conditions, allows accumulation of water (permanent or temporary), gives rise to a characteristic kind of soil and organisms adapted to these conditions and establishes dynamics coupled and interacting with economic and sociocultural flows that operate around and at different scales”. They also considered the wetland complexes as an ecologic unit constituted by a mosaic of diverse and contiguous wetlands in the landscape. Submarine areas were excluded from this analysis.

The categories considered were those proposed by Jaramillo et al. (2015), classifying wetlands based on the temporality of the water bodies according to the following attributes: permanent, temporary, open, canopy flooding, and land cover. To classify temporary wetlands, 13 categories from the Colombia National Map of Land Coverage and Use with Corine Land Cover legend were included, at a scale of 1:100 000 (IDEAM, 2016). In addition, as the biotic criterion, forests were differentiated into intertidal forests corresponding to mangroves, and alluvial forests, for a total of 15 analyzed covers (Table 1).

Table 1 Cover units used for the identification of the Colombian Continental Caribbean wetlands. Both the organization in four hierarchical levels. *Coverage denominations according to the National Land Cover Legend (IDEAM, 2010). ** Denomination of categories belonging to this study. 

The building of the input started by establishing a spatial framework that includes the coastal wetlands based on the delimitation criteria for the Colombia Coastal Environmental Units (UAC) (Alonso et al., 2003): 1) 2 km strip from the edge of the mangrove, 2) 100 % of the mangrove forests, 3) 2 km strip from the maximum elevation line of coastal lagoons; 4) all the areas of the National System of Protected Natural Areas bordering with the coastline, and 5) all the coastal towns within 2 km between the coastline and the edge of the urban perimeter. In addition, as an additional criterion, coastal geomorphological features were included (Figure 1A).

45 satellite Landsat 8 - OLI (available on https://earthexplorer.usgs.gov/) images, distributed in nine areas (Path/Row:7/51; 7/52; 8/51; 8/52; 9/52; 9/53; 10/53; 10/54; 10/55) and dated between March 2013 and May 2016 were identified to differentiate between temporary or permanent wetlands. Despite images include rainy and dry seasons in the Caribbean east and north areas due to El Niño phenomenon, dry conditions predominated mainly between March 2015 and February 2016 (UNGRD, 2016). These images were selected because they showed cloud cover under 10 % on the sites of interest, so its number changed from four to nine depending on the area, being higher in areas with drier weather.

To verify the image displacement, coordinates of known places were used and a non-significant displacement was determined, so no type of correction was applied. Radiometric calibrations were carried out through the FLAASH (Fast Line of Sight Atmospheric Analysis of Spectral Hypercubes) method included in the program ENVI that corrects the atmospheric effects on the images’ spectral response (Guo and Zeng, 2012).

After masking ocean pixels with cloud presence and their respective shadows, calculation of the NDVI (Normalized Difference Vegetation Index) index was carried out, which involves red and infrared bands, identifying flooded areas and borders between land and water (Rodríguez, 2001; Salinas et al., 2002; Lymburner et al., 2007; Zoffoli et al., 2007; Borro et al., 2009) with the following equation: NDVI = (ρIRC - ρR) / (ρIRC+ ρR). ρIRC is the near-infrared reflectance and ρR the red reflectance. This index normalizes values between -1 and 1, with a usual range between -0.75 to +0.75, where lower than zero values correspond to water bodies and greater than zero values to land cover (Thiam and Eastman, 1999).

A water/land threshold was identified in every image resulting from the NDVI calculation, with values between 0 and 0.06. For this purpose, a decision and information tree was used about the water bodies in the Colombia base cartography (IGAC, 2014). Starting from these thresholds the binarized segmentation was carried out, where every image was summarized in pixels with values 1 for water bodies and 0 for other coverages. Images were added to obtain time series to evidence the number of times each pixel detected the presence of water (Figure 1B), allowing to identifying permanent water bodies when their presence was detected at least 75 % of the times.

To complement the remote sensing processes, an analysis was carried out through thematic superposition of variables (Buzai and Baxendale, 2001) based on preexistent cartographic attributes (Table 2). Two uncertainty models were obtained, one for permanent wetlands and another for temporary wetlands, which provide information about the uncertainty of presence or absence of wetlands; according to the number of times the information was superimposed, an approximation about the presence and location of each wetland was achieved (Figure 1C and 1D). In addition, a model of flood frequency with 50 m resolution was used, based on images ALOS PALSAR 1, corresponding to seven detections between the years 2007 and 2011 (Quiñones et al., 2015).

Table 2 Attributes used in uncertainty models to identify permanent and temporary water bodies. The cartographic inputs used are: 1. IGAC Base cartography IGAC (IGAC, 2016), 2. National Map of de Ecosystems. (IDEAM et al., 2007; IDEAM et al., 2015), 3. Project GEF-SAMP, 4. Coastal erosion in the Colombian Caribbean (Posada and Henao, 2008), 5. Geopedology of Colombia (IGAC, 2014), 6. Colombia’s Land Coverage and Use (IDEAM, 2016), 7. Integral Diagnosisi of Coastal Lagoons (Rojas, 2014), 8. Colombia’s inland wetlands (IAvH, 2015; Jaramillo et al., 2015), 9. Colombia’s Mangroves -SIGMA (INVEMAR-MADS, 2016). 

Based on the multi-temporary method NVDI corresponding to a basic permanent and temporary water bodies input, manual editing was carried out focused on the low-certainty sites identified both in the uncertainty models and in the cited model of flood frequency, obtaining a layer of permanent and temporary water bodies. Later, the coverage attributes were assigned based on the cartographic layer of Colombia’s land coverage and use (IDEAM, 2016).

Both the manual editing and the assignment of coverage types were supported by 463 field-verified points. Points were distributed from Puerto Estrella in the north of La Guajira department, down to the south of the Uraba Gulf in the Antioquia department (Figure 1), most of them corresponding to high uncertainty sites. To capture the on-site information, directions RAMSAR (2010) were taken into account for quick ecological evaluations, with a capture time of 10 minutes per point, recording the following information: coverage percentage according to the vegetal structure, kind of vegetation, and physical characteristics of the landscape (flood traces, relief shape, and apparent hydric regime).

Once the product has been edited with the inclusion of the attributes, a vector union, and concatenation between permanent and temporary water bodies was carried out. Later, attributes of the soil moisture and the geomorphological feature by polygon were included according to the information of the Colombia geopedology layer at a scale of 1:100,000 (IGAC, 2014); this resulted in the cartographic product of the coastal wetlands of the Colombian Caribbean with a legend based on the temporality of the permanent and temporary water bodies, including the cover related to them.

Identification of wetland complexes and analysis of the wetlands cartographic product in the context of environmental management

Once the building of the cartographic product for coastal wetlands ended, it was used to identify the larger size wetland complexes and, based on them, to verify the contribution of the Protected Areas System to their protection. For this purpose, the categories of protected areas were included in the Colombia National Single Registry of Protected Areas - RUNAP (website inquiry on February 2020. http://runap.parquesnacionales.gov.co/). The size of the wetland types was calculated through Azimuthal Lambert projection using ArcGis 10.6 software.

Evaluation of the accuracy of the coastal wetlands cartographic product

To determine the accuracy of the cartographic product, field campaigns were conducted in four departments of the Colombian Caribbean with access and mobility facilities (La Guajira, Magdalena, Bolívar, and Sucre). The points distribution was randomly made in each legend class using the ArcGis tool Random Points, however, the points selected to visit were those easily accessible. At every visited point the type of water body observed was recorded in terms of the categories established in the cartography built. Likewise, the observed coverage was recorded in terms of the previously established categories in the cartographic product. To know the map’s total accuracy, an error matrix for each generated cartographic sub-product was built (water bodies and coverage). The total accuracy, the user accuracy, and the production accuracy were calculated based on it. Finally, the Kappa statistical analysis was applied, which provides a measure of the map accuracy degree, based on the hits recorded in the error matrix, and the theoretical success chance (Cohen, 1960). The value obtained with the Kappa statistics was interpreted according to the classification proposed by Landis and Kochy (1997), where values range from 0 and 1, concordance strength is good from 0.6 to 0.8, and it is deemed acceptable when it is greater than 0.81. To calculate the index the vcd - CRAN pack of the R program version 3.5.1 was used. (https://cran.r-project.org/web/packages/vcd/index.html).

RESULTS

What is the classification of coastal wetlands in the Colombian Caribbean and how are they geographically distributed?

576,627 ha coastal wetlands were delimited, which represent approximately 1.9 % of the Colombian wetlands. The scale of the concordance strength measured by kappa coefficient both for the temporality of water (k = 0.63) and for coverage (k = 0.79) was considerable or substantial.

Permanent wetlands covered 42.7 %, with permanent open bodies of water having the greatest relative coverage of 55.5 %, followed by intertidal forests with 32.2 % (Figure 2A). Most of the temporary wetlands were identified as settled on tidal flats or flood flats of alluvial origin, and to a lesser proportion on terraces, fans, and valleys. These wetlands covered 57.3 % of the total and were distributed in 15 types of coverage, of which herbaceous and/or bushy vegetation (30.9 %) and marshy areas (17.5 %) have the greater relative coverage. Rice fields, urbanized areas, and aquaculture ponds were the less common identified coverage in wetlands (Figure 2B).

Among the eight departments of the delimited coastal area, Magdalena has the largest coverage of coastal wetlands, followed by Antioquia; on its part, La Guajira, despite it is a desert, occupied third place. The fourth place was for the Caribbean area of Chocó, while departments Sucre, Córdoba, Bolívar, and Atlántico showed the lowest coverage (Figure 2C).

Figure 2 Classification of the Colombian Caribbean wetlands. A Percentage of permanent wetlands relative coverage. B Percentage of temporary wetlands relative coverage. C Percentage of permanent and temporary wetlands coverage respect to the total in the eight departments of the Colombian Caribbean. 

What are the most important coastal wetland complexes in the Colombian Caribbean and their types of coverage?

According to their size, the most important coastal wetland complexes in the Colombian Caribbean covered 498,740 ha, as follows: (i) Large Marsh of Santa Marta (235,556 ha), located in the Magdalena department bordering the Atlantic department on its east side; (ii) Urabá-Bajo Atrato (151,237 ha), located in the Caribbean south-western end, flanked by the Serranía del Darién on the west side and by the Serranía de Abibe on the east side; (iii) The Canal del Dique with 33,840 ha, located between departments Sucre and Bolivar, which covers a large part of the coastal area of the Barbacoas Bay; (iv) Cispatá Bay - Lower Sinú (33,401 ha) in the Córdoba department, on the fluvial-marine plain associated with the Cispatá Bay, Lower Sinú, and the delta formed by its outfall; (v) The Guajira’s salt flats located in the nort-western end of Colombia in the La Guajira department, with the largest one located in the area called middle Guajira, occupying 24,377 ha from the southwest of the Ranchería River’s delta and bordering 130 km along the coastal area down to the Cabo de la Vela in the north; other relevant desert wetland complexes in La Guajira are Bahía Portete (10,904 ha), and Bahía Honda and Hondita (9,422 ha) (Figure 3).

Figure 3 Location of the coastal wetland complexes in the Colombian Caribbean 

Regarding the type of wetland, the Ciénaga Grande de Santa Marta was the most diverse in types of coverage (n=16) identified, while Bahía Honda and Hondita complex was the least diverse with six types. The permanent canopy wetlands in the floodable intertidal forest were the largest type of wetland in the Cispata Bay (9,273 ha) and Canal del Dique (14,221 ha), while the open area temporary ones with little or no vegetation were the most important in Honda Bay and Hondita Bay (3,363 ha), and Portete Bay (3,778 ha). The permanent open wetland was the most important coverage in the Large Marsh of Santa Marta complex (99,617 ha), while the Temporary one in Herbaceous Vegetation (65,293 ha) and temporary in Salitral (9,531 ha) were the dominant ones in the Urabá Lower Atrato and La Guajira salt flats complexes, respectively. 77,539 ha of wetlands were identified outside the large Caribbean complexes (Table 3).

Table 3 Types of coverage in the Colombian Continental Caribbean Wetlands. Reference is made to the size of the coverage in each of the wetland complexes. BCBS: Cispata Bay and Lower Sinú; BHBH: Honda Bay - Hondita Bay; BP: Portete Bay; CD: Canal del Dique; CGSM: Large Marsh of Santa Marta; SGM: MIddle La Guajira Media Salt Flats; UBA: Urabá- Lower Atrato. N/A: Coverage identified outside any wetland complex. * Wetlands with evidence of transformation by use and artificial. 

Of all the mapped wetlands, 520,009 ha are deemed as natural, and 56,267 ha as transformed. Temporary wetlands in heterogeneous agricultural areas were the largest artificial wetland, while temporary ones in ponds for aquaculture were the smallest (Table 3).

How do the Caribbean Coastal Wetlands are represented in the National System of Protected Areas?

26.6 % of all the identified wetlands were found protected under some of the categories included in RUNAP. At the regional level, the Districts of Integrated Management were the figures that harbored most wetlands, while at the national level, The National Natural Parks did. Wildlife sanctuaries harbored the lowest part of wetland (Figure 4).

Figure 4 Contribution of the System of Protected Areas to the protection of the Colombian Caribbean Coastal Wetlands 

The Cispata Bay- Lower Sinú complex has the largest proportion of protected wetlands, 69.1 %, under the figure of the District of Integrated Management, followed by Portete Bay with 23.7 % under the figure of National Natural Park; the wetland complex of Urabá - Lower Atrato is protected to 23.1 % thanks to the Regional District of Integrated Management DRMI Blue Lake Los Manatíes and the Regional Natural Park (RNP) Wetlands of Rivers León and Suriquí; the complex Large Marsh of Santa Marta, protected to 23 %, distributed in two figures of the System of National Natural Parks: Fauna and Flora Sanctuary (FFS) Ciénaga Grande de Santa Marta and Salamanca Island Park Way (PW); on its part, La Guajira’s salt flats are protected to 15.3 % distributed the District of Integrated Management Musichi and DRMI Delta del Río Ranchería; Canal del Dique occupies the penultimate place, protected to 12.6 % by the PNN Corchal Mono Hernández and last, the complex Honda Bay and Hondita Bay which at the time of this publication had no protection category. It is noted that, although the Reserves of the Civil Society (RNSC) were not found among the big wetland complexes identified, they contributed with about 0.2 % of the protected area in the Caribbean in small disperse wetlands (Sanguaré, Hacienda, El Cequion, La Esperanza, Rivello, and Vigo). In general, the most part of the coastal wetland complexes are not found under protection figures (Figure 5).

Figure 5 Wetland complexes of the colombian caribbean and their protection figures. 

DISCUSSION

Classification and distribution

The identification of the biophysical limit of the Colombian Caribbean coastal wetlands shown here was mainly based on the determination of the water bodies in temporary or permanent, considering the flood pulse concept, where the wetland boundary changes in the floodplains between dry and rainy seasons, according to the description by Junk et al. (1989). However, the identification of limits in sites with high uncertainty was supported by biophysical features such as the type of coverage. On this matter, Cortés-Duque and Estupiñán-Suárez (2016) argue that vegetation is a good indicator of the transition area between the wetland and the land environment, although identifying the hydrophyte, helophyte, and hygrophyte vegetation is feasible only in the local delimitation or in small wetlands; also, there is not always vegetation as it happens in desert and semi-desert wetlands of the Caribbean coast; for the case of this cartographic product, which is at the regional level, using secondary information as the land coverage information (IDEAM, 2016) contributed to identifying wetlands by the type of vegetal coverage but also to identify urbanized wetlands and permanently free of vegetation areas such as the coastal salt flats. So, the Hydric-Coverage-Soil-Geomorphology approach was deemed best fit than Hydric-Vegetation-Soil-Geomorphology approach taking into account that the latter, like ephemeral wetlands, where there are contrasts appear between puddles and drought, foster that vegetation alternates with typically land plants (Johnson and Rogers, 2003).

The representation of wetlands in this cartography is different from the classification of wetlands in Colombia by Ricaurte et al. (2019) in that the flood pulse concept is explicitly reflected when differentiating wetlands by permanence or temporality of the water bodies. The representation of wetlands in this cartography differs from the classification of wetlands in Colombia generated by Ricaurte et al. (2019) in that the flood pulse concept which is explicitly reflected when differentiating wetlands by permanence or temporality of the water bodies. On its part, Ricaurte et al. (2019), differentiates wetlands with names that refer to the types of ecosystems. Because species do not distribute homogeneously and there are distribution gradients in many cases (Mumby et al., 1997; Ramirez, 1999), drawing boundaries at the level of ecosystems involves greater uncertainty both in their identification and delimitation; on the other hand, the coverage-based approach used to build the cartographic product presented in this publication, involves a lower degree of uncertainty and therefore it is adequate for monitoring transformation processes brought about by natural or anthropic drivers.

When separating the coastal from the inland wetlands, the generated cartographic product provides technical support to managing, planning, and ordering strategies, in the face of the climatic change perspectives described by IPCC (2019) for the coastal area, according to which, flood and erosive processes are expected in many areas of the world. Despite the described qualities, and considering that various wetland complexes extend into the country and show structural connectivity, it is relevant to integrate the cartographic product with that of the Colombia inland wetlands (Jaramillo et al., 2015). This work is possible because both show the same cartographic scale and compatibility in legends regarding the attribute of the temporality of the water bodies.

The contribution of the cartographic product presented here concerning Jaramillo’s et al. (2015) is significant because unlike the quoted authors, the building process includes field information and cartographic uncertainty models, which include the quoted Jaramillo’s et al. (2015) product; uncertainty models, as well as being replicable, are relevant in sites with low certainty about the presence of wetlands where the access to primary information is limited. The cartographic product is different from that by Jaramillo et al. (2015) because it contains information on five subcategories of permanent wetlands, and 15 subcategories of temporary wetlands.

Losing wetlands, besides affecting ecological processes of global importance, shows impacts on human development and wellbeing by increasing poverty and unemployment, and reducing opportunities for the sustainable development of communities (Ricaurte et al., 2019); to decrease this loss it is convenient to analyze opportunities of restoration and conservation for these ecosystems. If the delimitation of wetland complexes that are relevant for the Colombian Caribbean identified here, and their low protection level (Figure 4) are taken into account to ensure the supply of ecosystem goods and services, it is advisable to assess the possibility of expanding or creating new protected areas that favor the connectivity inside the identified complexes and allows to face global climate change scenarios under a sea-coast-inland connectivity approach. Given the heterogeneity of the identified coverage in the wetlands and its high supply of goods and services for the coastal communities, the Districts of Integrated Management were the most appealed protection figures in the Caribbean wetlands (Figures 4 and 5) and those that achieved the highest level of protection (e.g. 69.1 % of complex Cispatá Bay of- Lower Sinú), they could be a viable option for future declarations according to the different uses identified in the wetlands (agricultural, fishing, commercial, mining) (Table 2) and their potential to regulate the rational use of the natural resources and the environment. The absence of a protection category in the Honda Bay- Hondita Bay complex highlights the importance of prioritizing and keep making progress on the recent regional efforts to succeed in the declaration of this sector as a protected area (CORPOGUAJIRA, 2016).

Likewise, the declaration of international important areas (RAMSAR), is an option for the conservation and rational use of large areas of wetland complexes both natural and artificial (Ramsar Convention Secretary, 2016). This includes the CGSM estuary complex that became in its integrality covered as RAMSAR place, currently declared as such (even with higher coastal limits than those identified in this research) so that the cartographic product presented here is an additional input to update the Management Plan of the RAMSAR CGSM site, currently ongoing activity by the national and local authorities.

CONCLUSIONS

The first cartographic representation of exclusively coastal wetlands of the Colombian Caribbean was obtained, built up with the spatial reference system for Colombia Magna - Sirgas at a scale of 1:100,000, under a methodological approach based on analysis, interpretation, and integration of various spatial inputs that contributed to identify and delimit a product that could be integrated to the tools generated for the Colombia inland wetlands, by taking into account the connectivity of some of these coastal complexes with the continental ones.

The cartographic legend was based on a representation approach based on the temporality of the water and differentiating limits at the coverage level.

The classifications and descriptions presented from the generated cartographic product contribute to increasing the knowledge of seven wetland complexes of importance for the Colombian Caribbean: Large Marsh of Santa Marta (CGSM), Urabá-Lower Atrato, Cispatá - Lower Sinú, Canal del Dique, La Guajira’s Salt Flats, Portete Bay and Honda Bay - Hondita Bay. It was also identified that 74 % of the Colombian Caribbean coastal wetlands are not included in the conservation figures registered in RUNAP, so much of these figures leave the integrality of the seven big wetland complexes identified out. The cartographic product serves as an input to support the inclusion of new coastal wetlands under conservation figures that allow the departmental governments to make the most of their ecosystem services.

The methodology used to produce the wetlands cartography is replicable, and together with the wetlands classification, provides inputs to fill the gaps identified by RAMSAR convention about data and information needs for the rational use and management of coastal areas and designation of new sites.

ACKNOWLEDGMENTS

The authors would like to thank the Ministry of Environment and Sustainable Development (Minambiente) for fostering the development of this research, Paola Sáenz for accompanying the process from Minambiente and Julián Pizarro for accompanying the process from INVEMAR; likewise, work of Liliana Barreto, Diana Romero, David Forero, and Stephan Rivera are acknowledged, who contributed with their work to build the cartographic product.

BIBLIOGRAFÍA / LITERATURE CITED

Alongi, D. 2015. The impact of climate change on mangrove forests. Curr. Clim. Change Rep., 1:30-39. doi: 10.1007/s40641-015-0002-x. [ Links ]

Alonso, D., P. Sierra-Correa, F. Arias-Isaza y M. Fontalvo. 2003. Conceptos y guía metodológica para el manejo integrado de zonas costeras en Colombia, manual 1: preparación, caracterización y diagnóstico. Serie Documentos Generales INVEMAR. No. 12, 94 p. [ Links ]

Andrade-Amaya, C. 2001, Las corrientes superficiales en la Cuenca de Colombia observadas con boyas de deriva. Rev. Acad. Col. Cienc. Ex., Fís. Nat., 96: 321-335. [ Links ]

Blanckespoor, B., S. Dasgupta and B. Laplante. 2014. Sea-level rise and coastal wetlands. Ambio, 43: 996-1005. [ Links ]

Borro, M., M. Salvia, P. Minotti, A. Puig, H. Karzenbaum y P. Kandus. 2009. Primeros resultados de la clasificación de las lagunas someras en la clasificación del Delta del Paraná bajo un enfoque ecohidrogeomorfológico. Jornadas argentinas ecología paisaje. 91-104. [ Links ]

Buzai, G. y C. Baxendale. 2001. Crecimiento urbano y pérdida de suelos productivos. Análisis espacial mediante tecnología SIG. Ger. Amb., 9: 697-701. [ Links ]

Cohen, J. 1960. A coefficient of agreement for nominal scales. Educ. Psychol. Meas., 20: 37-46. [ Links ]

CORPOGUAJIRA. 2016. Socialización ante la comunida del estudio de línea de base para declaratoria de áreas protegidas bahía Honda y bahía Hondita. Tomado de http://corpoguajira.gov.co/wp/corpoguajira-y-la-fundacion-george-dahlLinks ]

Cortés-Duque, J. y L. M. Estupiñán-Suárez. (Eds.) 2016. Las huellas del agua. Propuesta metodológica para identificar y comprender el límite de los humedales de Colombia. Inst. Invest. Rec. Biol. Alexander von Humboldt, Fondo Adaptación. Bogotá. 340 p. [ Links ]

Davidson, N., E. Fluet-Chouinard and C. Finlayson. 2018. Global extent and distribution of wetlands: trends and issues. Mar. Freshw. Res., 69: 620-627. [ Links ]

Estupiñán-Suárez, L. 2019. A classification system for Colombian wetlands: an essential step forward in open environmental policy-making. Wetlands, 39: 1-20 [ Links ]

Giri. C., E. Ochieng, L. Tieszen, Z. Zhu, A. Shingh, T. Loveland, J. Masek and N. Duke. 2011. Status and distribution of mangrove forests of the world using earth observation satellite data. Global Ecol. Biogeogr., 20: 154-159. [ Links ]

Guo, Y. y F. Zeng. 2012. Atmospheric correction comparison of SPOT 5 image based on model FLAASH and model QUAC. Internat. Arch. Photogr. Rem. Sens. Spat. Inf. Sci., 39-B7, XXII ISPRS Congress, Melbourne. [ Links ]

HIMAT. 1988. Calendario meteorológico. Inst. Col. Hidrol. Meteorol. Adecuac. Tierras, Bogotá. 227 p. [ Links ]

IAvH. 2015. Productos cartográficos de identificación de humedales; tipos de humedales y transformación de humedales. Convenio marco 005 (13-014) entre Inst. Alexander von Humboldt y Fondo Adaptación. [ Links ]

IDEAM, IGAC, IAvH, INVEMAR, SINCHI. 2007. Ecosistemas continentales, costeros y marinos de Colombia. Inst. Meteor. Est. Amb., Inst. Geogr. Agustín Codazzi, Inst. Invest. Rec. Biol. Alexander von Humboldt, Inst. Invest. Amb. Pac. John von Newmann, Inst. Invest. Mar. Cost. José Benito Vives de Andreis, Inst. Amaz. Invest. Cient. Sinchi. Bogotá. 276 p. + 37 hojas cartográficas. [ Links ]

IDEAM. 2010. Leyenda nacional de coberturas de la tierra. Metodología CORINE Land Cover adaptada para Colombia Escala 1:100.000. Inst. Hidrol., Meteor. y Est. Amb. Bogotá. 72 p. [ Links ]

IDEAM. 2012. Mapa de clasificación climática de Caldas - Lang. Inst. Hidrol., Meteorol. Est. Amb. - IDEAM, Cartografía Básica [ Links ]

IGAC. IDEAM, MADS, IAvH, SINCHI, INVEMAR, INAP, PNNC e IGAC. 2015. Mapa de ecosistemas continentales, costeros y marinos de Colombia a escala 1:100.000. Convenio 4206 de 2011. [ Links ]

IDEAM. 2016. Mapa nacional de cobertura y uso de la tierra a escala 1:100000. Inst. Hidrol., Meteorol. Estud. Amb., Bogotá. [ Links ]

IGAC. 2014. Mapas de geopedología del territorio colombiano a escala 1:100.000, formato shapefile. Inst. Geogr. Agustín Codazzi. Bogotá. [ Links ]

IGAC. 2016. Cartografía base de Colombia a escala 1:100.000. Formato shapefile. Inst. Geogr. Agustín Codazzi. Bogotá D.C. [ Links ]

INVEMAR-MADS 2016. Resolución No. 478 de 2016. Componente 2. 148 p + Anexos. [ Links ]

INVEMAR. 2019. Informe del estado de los ambientes y recursos marinos y costeros de Colombia, 2019. Serie Publ. Periód. 3, 183 p. [ Links ]

IPCC. 2019. Special Report on the Ocean and Cryosphere in a Changing Climate [H.-O. Pörtner, D.C. Roberts, V. Masson-Delmotte, P. Zhai, M. Tignor, E. Poloczanska, K. Mintenbeck, A. Alegría, M. Nicolai, A. Okem, J. Petzold, B. Rama, N.M. Weyer (eds.)]. In press. [ Links ]

IPCC. 2019. Special report on the ocean and cryosphere in a changing climate. In: Pörtner, H.-O., D.C. Roberts, V. Masson-Delmotte, P. Zhai, M. Tignor, E. Poloczanska, K. Mintenbeck, A. Alegría, M. Nicolai, A. Okem, J. Petzold, B. Rama and N.M. Weyer (Eds.) [ Links ]

Jaramillo, U., J. Cortés-Duque y C. Flórez (Eds). 2015. Colombia Anfibia. Un país de humedales. Vol. 1. Inst. Invest. Rec. Biol. Alexander von Humboldt. Bogotá. 140 p. [ Links ]

Johnson, P. and G. Rogers. 2003. Ephemeral wetlands and their turfs in New Zealand. Sci. Cons., 230. 109 p. [ Links ]

Junk, W., J. P. Bayley and R. Sparks. 1989. The flood pulse concept in river-floodplain systems. Can. Spec. Publ. Fish. Aq. Sci., 106: 110-127. [ Links ]

Landis, J. and G. Koch. 1977. The measurement of observer agreement for categorical data, Biometrics. 159-174. [ Links ]

Lotze, H., H. Lenihan, B. Bourque, R. Bradbury, R. Cooke, M. Kay, S. Kidwell, M. Kirby, C. Peterson and J. Jackson. 2006. Depletion, degradation and recovery potential of estuaries and coastal seas. Science, 312: 1806-1809. [ Links ]

Lymburner, L., D. Burrowa and B. Butler. 2007. Using remote sensing to map wetland water clarity and permanence: approaches for identifying wetlands requiring management in large catchments. Proc. 5th Int. Australian Stream Management Conference. Australian rivers: making a difference, 235-240 [ Links ]

Mancera, J., W. Troncoso, V. Rivera-Monroy, R. Twilley y R. Giraldo. 2001. Variaciones estacionales e interanuales durante los últimos 30 años en la calidad de las aguas de la Ciénaga Grande de Santa Marta: una laguna costera tropical de ambiente árido. IX Congr. Latinoam. Cienc. Mar, San Andrés. [ Links ]

Mumby, P., E. Green, A. Edwards and C. Clark. 1997. Coral reef habitat mapping: how much detail can remote sensing provide? Mar. Biol., 130: 192-202. [ Links ]

Posada, B y W. Henao. 2008. Diagnóstico de la erosión en la zona costera del Caribe colombiano. Serie Publ. Espec., 13, Invemar. 200 p. [ Links ]

Quiñones, M., M. Vissers, A.M. Pacheco-Pascaza, C. Flórez, L. Estupiñán y C. Aponte. 2015. Un enfoque ecosistémico en el análisis de una serie densa de tiempo de imágenes de radar Alos PALSAR, para el mapeo de zonas inundadas en el territorio continental colombiano. Aplicaciones en el mapeo de frecuencias de inundación como insumo al mapeo de humedales. Biota Col.,17 (Suplem. 1): 63-85. Diseño y análisis estadístico. Univ. Jorge Tadeo Lozano, Bogotá. 325 p. [ Links ]

Ramírez, A. 1999. Ecología aplicada: Diseño y análisis estadístico. Univ. Jorge Tadeo Lozano, Bogotá. 325 p. [ Links ]

RAMSAR. 2010. Directrices para la evaluación ecológica rápida de la biodiversidad de las zonas costeras, marinas y de aguas continentales. Secret. Conv. Divers. Biol., Montreal, Serie Publ. Tecn. 22, Inf. Técn. Ramsar 1. 54 p. [ Links ]

RAMSAR. 2016. Introducción a la Convención sobre los Humedales, Secretaría de la Convención de Ramsar, Gland, Suiza. 116p. [ Links ]

RAMSAR, 2018. Convención de Ramsar sobre los Humedales. Perspectiva mundial sobre los humedales: Estado de los humedales del mundo y sus servicios a las personas. Gland (Suiza). Secretaría de la Convención de Ramsar. 88 p. [ Links ]

Ricaurte, L., J. Patiño, D. Zambrano, G. Arias, O. Acevedo and C. Aponte. 2019. A classification system for Colombian wetlands: an essential step forward in Open Environmental Policy-Making. Wetlands. 39: 971-990. [ Links ]

Rocha. J., J. Yletyinen, R. Biggs, T. Blenckner and G. Peterson. 2015. Marine regime shifts: drivers and impacts on ecosystem services. Phil. Trans. R. Soc. B., 370, 20130273. [ Links ]

Rodríguez, I. 2001. Comparación de técnicas basadas en el tratamiento digital de imágenes de satélite para la obtención de la línea de costa del Delta del Ebro. Teledetección Medio Ambiente Cambio Global, 205-208. [ Links ]

Rojas, X. 2014. Diagnóstico integral de las lagunas costeras: Documento fase II. Convenio 156 MADS- ASOCARS. Bogotá, 112 p. [ Links ]

Salinas, C., W. E. Treviño, J. Jaramillo y F. Campos. 2002. Identificación y clasificación de humedales interiores del estado de Tamaulipas por percepción remota y sistemas de información geográfica. Invest. Geogr.,74-91. [ Links ]

Thiam, A., and J. Eastman. 1999. Vegetation Indices, Guide to GIS and Image Processing. Vol. 2. Clarks Lab. Clark University. [ Links ]

UNGRD, 2016. Fenómeno El Niño. Análisis comparativo 1997-1998/2014-2016. Un. Nal. Gest. Riesgo Desast, Presidencia República. 142 p. [ Links ]

Vilardy, S., U. Jaramillo, C. Flórez, J. Cortés-Duque, L. Estupiñán, J. Rodríguez, O. Acevedo, W. Samacá, A. Santos, S. Pelaez y C. Aponte. 2014. Principios y criterios para la delimitación de humedales continentales: una herramienta para fortalecer la resiliencia y la adaptación al cambio climático en Colombia. Inst. Invest. Rec. Biol. Alexander von Humboldt. Bogotá, 100 p. [ Links ]

Worm, B., E. Barbier, N. Beaumont, J. Duffy, C. Folke, B. Halpern, J. Jackson, H. Lotze, F. Micheli, S. Palumbi, E. Sala, K. Selokoe, J. Stachowicz and R. Watson. 2006. Impacts of biodiversity loss on ocean ecosystem services. Science, 314: 787-790. [ Links ]

Zoffoli, M., N. Mabanes y P. Kandus. 2007. Contribución de series temporales de NDVI NOAA/AVHRR al análisis funcional en humedales: 6299-6306. Anais XIII. Simp. Brasil. Sensoriamento Remoto, Florianópolis, Brasil. [ Links ]

Received: February 12, 2020; Accepted: March 07, 2020

*Autor de correspondencia: santiago.millan@invemar.org.co

Creative Commons License This is an open-access article distributed under the terms of the Creative Commons Attribution License