Dieback of mangroves in the Jaguaribe River Estuary, Northeastern Brazil and the associated anthropogenic drivers

Abstract

Mangroves are threatened by direct (shrimp aquaculture, urbanization) and indirect (climate change) drivers, which act synergistically in coastal settings causing degradation and loss. In NE Brazil these drivers strengthened in recent decades. A 25-year series of remote sensing images of land use changes in the Jaguaribe River Estuary, with emphasis on mangroves, showed a mangrove dieback of 41% relative to the 2000 cover (803 ha). In this period mangrove area increased from 2000 to 2009, reaching 851 ha (+4.8 ha.yr-1). Thereafter loss rate was much higher (-37.2 ha.yr-1), reducing mangrove cover to 479 ha in 2023. Drivers of dieback were the increase (161%) of shrimp farming area and urbanization (119%) since 2000. A positive correlation (p < 0.01) between urbanization and shrimp farming expansion, suggests that infrastructure supporting shrimp farms (roads, bridges and wind power fields) are a significant pressure. Direct conversion of mangroves into shrimp ponds was minor, but a time-delayed effect on mangrove health bordering aquaculture activities was observed. These drivers also caused a reduction of salt flats (‘apicuns’). The impact of shrimp farming and its associated infrastructure on mangroves is overlooked, since no direct forest conversion is evident, rather a time-delayed continuous loss of ecosystem integrity occurs.

Key words
land use change; estuary; semiarid coast; shrimp aquaculture; El Niño

INTRODUCTION

Globally, mangroves occupy around 147,359 km2 of sheltered tropical and subtropical coasts of river deltas, estuaries and other sheltered coastal settings (Bunting et al. 2022). These ecosystems are threatened by several direct (shrimp aquaculture, conversion to oil palm or other commodities, urbanization) and indirect (climate change) anthropic threats. Synergy of both types of threats can yet deepen mangrove loss, which, despite at a lower rate, continues worldwide.

Mangrove forests are ubiquitous along a great part of the 7,367 km of Brazilian coast, with their most extensive areas in the north coast, up to 3° Latitude, with 70-75% of the country coverage (Lacerda et al. 2022). Mangrove extension along the Northeastern Brazilian coast responds for less than 5% of the total Brazilian cover of these forests, of about 11,100 km2. However, in this mostly semiarid, mesotidal coastline, where continental runoff is minimum and coastal erosion is ubiquitous, mangroves show limited distribution in narrow bands along ephemeral estuaries subjected to long dry periods and multi-year droughts. Notwithstanding, despite their restricted relative extension, northeast Brazil mangroves provide a wide range of ecosystem services, mostly associated with coastal protection, carbon sequestering and stocking, and traditional fisheries, that provide vital resources for a growing piece of population that rely on mangrove resources in this region (Lacerda et al. 2019, 2024).

Unfortunately, the NE mangroves of Brazil are one of the most impacted ecosystems because combined influences of climate change and local anthropogenic drivers, leading to conversion and loss, are already ongoing (Lacerda et al. 2024). A major driver of environmental impacts is intensive shrimp farming, since this sector of the Brazilian coast responds to over 96% of cultivated shrimp production in Brazil (Lacerda et al. 2021). Although seldom directly leading to deforestation, shrimp farming alters significantly the morphology, hydrochemistry and eutrophic state of estuaries negatively impacting mangrove health (Marins et al. 2020). In addition, recent changes of the Brazilian Forest Code (BFC) in 2012, have weakened protection for mangroves and associated salt flats (“apicuns”). Apicuns are mangrove-associated ecosystems that are part of a dynamic ‘continuum‘ of estuarine ecosystems, and home to an extremely ecologically fitted (e.g. Cyanophyta, fiddler crabs) but threatened biota (e.g. the crab Cardisoma guanhumi) (Schmidt et al. 2013, Ferreira et al. 2022). After extensive mangrove forest clearing in past decades by shrimp farming activities in northeast Brazil, some enforcements in monitoring over mangrove ecosystems in the last years have been implemented - in part due to the increase of concern about the ongoing and predicted effects of climate change over coastal areas. Hence, salt flats have been then the preferable area for shrimp ponds construction, rather than the mangrove coverage, but ponds keep periodic hydrological connectivity through dammed channels, allowing the flushing of effluents to estuaries (Meireles et al. 2007). We hypothesize that urban development is associated with increasing shrimp farming, such as opening and maintain of access roads and bridges and construction of production infrastructure also positively feedback with shrimp farming.

Several authors agree that the main impacts of shrimp farming on Brazilian mangroves presently are indirect, because of excess nutrients and sediment inputs from shrimp pond effluents resulting in loss of ecosystem services, such as reductions in primary productivity, carbon storage capacity, efficiency as estuarine filters, biodiversity and abundance of subsistence fisheries, and resistance to other environmental stressors (Queiroz et al. 2013, Lacerda et al. 2021). In addition, soil damage and infrastructure remaining after shrimp pond deactivation impairs mangrove recovery. This extends the duration of the damage and allows the occupation of degraded areas by other activities that can permanently impair ecosystem function (Ferreira and Lacerda 2016a). In this study, we address the shrimp culture farming increase in the Jaguaribe River estuary, the largest cultivated shrimp producer in Brazil, its interactions with other anthropogenic drivers and consequences to the future of mangroves in the region considering climate change and coastal development.

MATERIALS & METHODS

Environmental setting

The Jaguaribe river estuary (4° 26’ S and 37° 47’ W) (Figure 1), located in northeastern Brazil, is under a hot semiarid climate, with average annual rainfall of 794 mm, occurring mostly from January to May, and average temperatures varying between 26°C and 28°C. Between 2011 and 2017 (Figure 2), an extreme drought period lowered annual mean rainfall to 506 mm compared to the historical mean of 794 mm (FUNCEME 2025). The estuary has a length of 36 km, is under a semidiurnal tidal regime with highest amplitude of 2.8 m and occupies an area of approximately 3,050 km²; total freshwater fluxes vary from 22 to 57 m3.s-1 (Dias et al. 2009, 2011). At the upper reaches of the estuary a dam built at Itaiçaba village, artificially separates the estuary from the major portion of the lower Jaguaribe River basin (Figure 1).

Figure 1
Figures S1-s4

Figure 2
Major drivers of land use change in the estuarine basin in the Jaguaribe River, NE Brazil. Based on images taken in 2023 (Figure S1). Salt flats are narrow bands between mangroves and upland areas and are not shown in this low-resolution map.

The estuarine watershed displays a diversity of land covers and uses, including farmland, pastures, urban areas and original vegetation. The right margin of the estuary is dominated by sand dunes, whereas the left margin is dominated by cliffs, from the Tertiary Barreiras Formation. The catchment responds for about 70% of all freshwater that reaches Atlantic Ocean off the western portion of the northeastern equatorial coast of Brazil (Marins et al. 2002).

Major anthropogenic drivers of land use changes in the estuary are intensive shrimp aquaculture, presently covering about 2,200 ha in the middle and upper estuary, and responsible by c.a. 12% of the national shrimp production; two medium size cities (Aracati and Fortim) and several small villages, totalling about 100,000 inhabitants, with only partial treatment of wastewaters and urban solid wastes (Marins et al. 2007, 2011). Other anthropogenic activities are agriculture, husbandry and mining, all located outside the estuarine basin and without significant direct impact on mangroves (Figure 1). Another important driver of land use change are the presence of extensive dune fields and sandy beaches, which, due to the dominating semiarid conditions, are relatively mobile and can suffocated mangrove, as observed in other estuaries along the northeastern coast (Maia et al. 2005, Lacerda 2018, Lacerda et al. 2024). At the Jaguaribe estuary, aeolian sand transport was estimated to be about 200,000 m3.yr-1 (Marins et al. 2002). Yet, in the last few years, countless wind power fields have emerged in the region, with their associated infrastructure and activity.

The mangroves of the Jaguaribe river estuary are composed by Rhizophora mangle Avicennia schaueriana, A. germinans and Laguncularia racemosa. The species are mixed and show no zonation in most of the estuary, but in its higher reaches, where landward expansion is very clear, fringe forests are dominated by monospecific stands of L. racemosa. Mangroves and salt flats covered areas of about 803 ha and 103 ha respectively, in 2000, year of the beginning of the present analysis. Near the total of the area covered by these two physiognomic classes occur along the right margin dominated by fluvial and estuarine plains and show a very restricted distribution barred by the cliffs of the Barreiras Formation in the left margin (Figure 1).

Remote sensing procedures

Multitemporal analysis was performed using Landsat 5 TM (2000–2012) and Landsat 8 OLI (2013–2023) images, with a spatial resolution of 30 m. The images were obtained from the Google Earth Engine (GEE) platform and preprocessed in QGIS 3.28.15 using the Semi-Automatic Classification (SCP) plugin. Radiometric and atmospheric corrections were applied using the DOS1 and FLAASH models (Chander et al. 2009) to minimize atmospheric interference.

A combination of visual interpretation and supervised classification was used to identify land use and land cover classes. Visual interpretation was based on morphological and spectral criteria (shape, size, hue, and spectral pattern; Lillesand et al. 2015), supporting the delimitation of mangroves. The land cover and land use classes established by MapBiomas for the year 2023 were used: mangrove, aquaculture, agriculture, non-vegetated areas, water bodies (rivers, lakes, and ocean), forest formation, savanna formation, and natural non-forest formation. These categories allowed for a detailed analysis of land use and cover dynamics in the Jaguaribe River estuary, with an emphasis on identifying the extent and temporal variation of mangroves and the main anthropogenic vectors associated with their degradation. Salt flats were also recognized, although their narrow extent makes clear detection in Landsat images difficult.

Supervised classification was conducted in SCP using the Random Forest (RF) algorithm, trained with spectral signatures extracted from regions of interest (ROIs). Accuracy assessment was performed using stratified random sampling, with 70% of the samples used for training and 30% for validation. A confusion matrix was constructed to calculate user accuracy, producer accuracy, overall accuracy (OA), and the Kappa index (Congalton & Green 2019). The classification achieved an overall accuracy (OA) of 81%, a value considered satisfactory according to Congalton (1991), who suggests values above 0.80 as indicative of a good classification. The Kappa index had a value of 0.78, which, according to the interpretation of Landis & Koch (1977), corresponds to substantial agreement between the classification and the reference data.

Additionally, the consistency of the classifications was evaluated by comparison with the historical series of MapBiomas (Collection 8.0; MapBiomas 2023). The spatial analysis showed agreement of approximately >85% between the two products, with greater divergences observed in the exposed soil and secondary vegetation classes, reflecting differences in spatial scale and class definition. This comparison showed that, although MapBiomas represents a robust basis on a national scale, the classification developed in this study captured the local boundaries of mangroves and adjacent uses with greater accuracy.

Finally, the spatial quantification of land use and land cover classes was performed in the GRASS GIS environment using the r.report module. The results were expressed in hectares and as a percentage of the total estuary area. Positional uncertainty (±30 m) was considered based on the spatial resolution of Landsat images and georeferencing errors (Pogson & Smith 2015). Maps generated by the described procedures are provided as Figure S1.

ENSO effects

To assess global scale climate drivers’ eventual effects on our study area, we correlated the Ocean Niño Index (ONI) with the different land cover in the Jaguaribe Estuary, since El Ninõ has been shown to strongly affect mangrove dynamics in different regions of South America (Rebolledo Monsalve et al. 2024). ONI is an index for tracking ocean path of ENSO, calculated as the increased anomalous temperature from an average of the recorded temperature of 3 consecutive months of ocean surface waters; values ≥ of 0.5 indicate El Niño, values ≤ 0.5 indicate La Niña (NOAA 2025).

Multivariate data analyses

Multivariate data analyses were performed using principal component analysis (PCA) after data standardization. PCA is a widely used exploratory data analysis tool that identifies patterns and relationships between variables and observations. By plotting the loadings of these data’s principal components (PC1 and PC2), groups of samples with similar behavior are identified, as well as the existing association among the original variables. Pearson correlation coefficients were applied to confirm the relationship between drivers (variables) and land use changes. This analysis helped assess the strength and direction of the relationships between variables. All statistical analyses were set at a significance of p < 0.05 and performed using the software GraphPad Prism version 9.5.0 (La Jolla, California).

RESULTS

As evidenced by the historical series of land use cover (Figure S1), the estuarine basin of the Jaguaribe river experienced significant changes. Nearly half of the mapped area covered by upland forests, for example, was converted to savanna-like vegetation, mostly pasture, since no significant change in area dedicated to agriculture occurred in the 23-years period. Although there is high environmental significance for these changes, only land use changes directly associated with mangroves are presented in this study. Table I summarizes the changes in area of the different land use classes in the Jaguaribe River Estuary between 2000 and 2023, including natural features (mangroves, salt flats and dunes) and the major anthropogenic drivers of land use change located in the estuarine basin directly related to mangroves (shrimp farms and urbanization).

Table I
Changes in area (ha) of the different land use and land cover classes in the Jaguaribe River Estuary, NE Brazil, between 2000 and 2023.

Year	Mangroves	Salt flats	Dunes	Shrimp farms	Urbanized areas
2000	803	103	4,713	875	978
2003	818	121	4,727	1,388	1,112
2006	837	133	4,654	1,773	1,209
2009	851	117	4,620	1,992	1,265
2012	838	97	4,635	2,177	1,425
2015	785	111	4,547	2,314	1,537
2018	490	116	4,448	2,189	1,704
2022	522	82	4,374	2,465	1,817
2023	479	60	4,343	2,284	2,145

The historical data series shows a breaking point after 2006. The area covered by mangroves and salt flats increased steadily from 2000 to 2009 and 2000 to 2006, respectively. Whereas mangroves total extension increased from 803 ha in 2000 to 851 ha (+4.8 ha.yr-1) in 2009; salt flats area increased from 101 ha in 2000 to 133 ha in 2006 (+4.2 ha.yr-1). Both land cover classes decreased steadily thereafter. The magnitude of mangroves and salt flats loss rates (-37.2 ha.yr-1 and -6.1 ha.yr-1 relative to 2000 values, respectively), were much higher in this period than the gain rates observed from 2000 and 2009 and 2006 respectively. This resulted in 2023 in a net loss of mangrove and salt flat cover to 324 ha and 43 ha, respectively, about 41% reduction relative to 2000 extension, for both land cover classes.

An extensive area of the estuarine basin is covered by fixed and mobile dunes. The acceleration of dune field displacement has been observed along the NE Brazil coast, including in the Jaguaribe estuary, but its long-term displacement is, until now, not quantified (Maia et al. 2005). Our results show a decreasing trend from 2000 to 2009, prior to the onset of the last extended drought in the region, from 4,713 ha in 2000 to 4,343 ha in 2009, a 7.9% of total area reduction.

Contrary to mangroves and salt flats, shrimp farming and urban areas increased steadily throughout the entire period from 875 ha and 978 ha in 2000, to 2,284 ha and 2,147 ha in 2023, respectively. Average annual increase rates were +50.7 ha.yr-1 and +61.3 ha.yr-1, respectively, and resulted in total increase of shrimp farm area of 161% and of urbanized areas of 119% in 2023 relative to 2000. Most of this increase occurred over natural vegetation, but also including salt flats and mangroves (Figure S2). Whereas shrimp farms are spread along the entire estuary, mostly at the margin of rivers and tidal creeks of the left margin, where fluvial plains dominate, urban areas are concentrated in the four cities and villages located in the estuarine basin, three of them located in the right margin dominated by the Tertiary Barreiras Formation (Figure 1). The largest Aracati city, at the left margin, expands mostly inland on fluvial terraces and a straight coastal plain, with only about 2.5 km of its perimeter bordering the river, but other river front developments and infrastructure supporting extensive wind power fields, and access to shrimp farms seem to have affected mangrove areas (Figure S3).

Pearson correlation analysis of the measured variables (Table II) shows significant positive correlations between shrimp farming and urbanized areas (rs = 0.806; P < 0.01), suggesting that these drivers are strongly associated, mostly through access and logistics development necessary to cope with the increasing shrimp production in the region. On the other hand, urbanized area expansion and, to a lesser extent, shrimp farms are negatively correlated with natural land cover (mangroves, salt flats and dunes).

The expansion and reduction trends of these different natural land cover areas, therefore, suggest a response to one or more of the anthropogenic drivers of land use change, such as urbanization and shrimp farming locally, as suggested by the correlation matrix (Table II).

Table II
Spearman correlation matrix between land use change drivers in the Jaguaribe River estuary between 2000 and 2023. (ns = non-significant; ** p < 0,01; “n” is the number of image data used for the correlation: n = 9).

Driver	Mangrove	Shrimp farms	Salt flats	Urban area	Dunes
Mangrove	1
Shrimp farms	-0.514ns	1
Salt flats	0.635ns	-0.384ns	1
Urban area	-0.856**	0.806**	-0.745**	1
Dunes	0.911**	-0.772**	0.702**	-0.856**	1

The expansion of urbanized areas showed a strong negative and significant correlation (p < 0.01) with mangrove extension (Table II). In agreement with the correlation analysis, the evolution of land cover shows an expansion of urbanization related to mangrove degradation, and there is a direct relationship between urbanization expansion and mangrove dieback, whereas a time lag of a few years exists between shrimp pond expansion and mangrove dieback. Figure 3 depicts the velocity of area change of each land use. Local anthropogenic drivers are also accompanied by a decrease in extension of dune fields, which shows a significant negative correlation with shrimp farming and urbanization (Table II).

Figure 3
Annual rainfall in the Jaguaribe River Estuary, Aracati meteorological station (FUNCEME 2025). The continuous line refers to the mean of the past consecutive two years.

There is a direct relationship between urbanization and mangrove dieback, whereas a time lag of a few years exists between shrimp pond expansion and mangrove dieback. The strength of the El Niño (ENSO) events has been suggested to have a strong effect on coastal vegetation including mangroves and salt flats (“apincuns”) therefore we performed a correlation analysis between a well-accepted ENSO strength scale, the Ocean Niño Index (NOAA 2025) with mangrove area changes during the 2000-2023 period (Figure 4). The results showed no significant correlation between ONI and mangroves (rs = 0.535; ns) or salt flats (rs = 0.535; ns) areas, different of dune field area which, as expected, was significantly correlated with ONI (rs = 0.674; p < 0.05).

Figure 4
Evolution of mangrove, urban and shrimp pond areas in the Jaguaribe estuary between 2000 and 2023.

Multivariate data analysis

The PCA analysis of all variables of interest shows the interaction between different land uses and the degree of impact of the intensity of El Niño, as expressed by the ONI index (Figure 5). Shrimp farm area and urban areas clearly positively covary while the three natural land cover classes negatively vary with the local anthropogenic drivers. The longer arrow representing shrimp farms, suggest aquaculture with the stronger influence on mangroves and salt flats. Shrimp pond area and salt flats are closer and with high negative contribution of the vertical axis (PC2).

Figure 5
Ocean Niño Index (NOAA 2025) and mangrove area change (right y axis, black line) in the Jaguaribe River Estuary in NE Brazil.

As for the El Niño influence, the moderately negative correlation with PC1 (-0.460) suggests that years with El Niño (higher ONI) tend to be associated with lower PC1 values, expansion of natural environments such as mangroves and sand dunes. The weak correlation with PC2 (–0.224) indicates a less clear pattern, perhaps with some indirect influence on salt pans, but not strong enough. None of the models has statistical significance (p < 0.05), but mangroves and sand dunes show moderate positive trends with R² values of 0.25 and 0.23. The ONI index explains up to 25% of the variability in mangrove areas, which is ecologically relevant, even considering the lack of statistical significance due to the small sample, and feedback the significant negative impact of local drivers.

DISCUSSION

In 2012 Brazil promulged amendments to the Brazilian Forest Code (BFC) (created in 1934 and several times modified), with major impacts on mangrove forests and salt flats (apicuns); major changes included a narrowing of the protection strip of riparian vegetation along waterbodies and the release of salt flats to establish locally economic important activities, mostly aquaculture, tourism and urban development and the necessary infrastructure, including roads, harbours and bridges to access them. According to Ferreira & Lacerda (2016b) this opened over 600,000 ha of coastal plain areas and salt flats for development. Although mangroves remained as areas of permanent protection, these measures strongly threatened their integrity. It is important to note that the reduction in mangrove area did not always result from deforestation, but rather, due to dying of trees, as seen in the images collection in Figure S2, and this makes monitoring of affected areas difficult. The amendments of the BFC also lack any legislation regulating the water intake and spillage by shrimp ponds, which can directly affect forests through altering hydrology and salt exchange, soil erosion and tidal creeks clogging, and excess nutrient inputs. In addition, the amendments to the BFC did not promote any compensatory measurements related to mangrove, neither, at least, the removal of dykes in deactivated shrimp ponds and saltworks or to partially re-establish hydrological connection with the estuary, that would allow partial mangrove recovery. A thoroughly discussion of the general impacts of the revised Forest Code on Brazilian mangroves and other riparian systems can be accessed in Silva et al. (2012), Ferreira & Lacerda (2016a, b) & Oliveira Filho et al. (2016).

According to Lima et al. (2025) satellite data between 2008 and 2023, in the rivers of the East coast of Ceará State, including the Jaguaribe Estuary, but also three other estuarine regions Pacoti, Piranji, Manibú, salt flat and mangrove areas decreased in 50.02% and 16.76% respectively, while shrimp culture (together with minor salt pond areas) increased 35.08%. The observed shift from expansion to decrease of mangrove and salt flat areas after 2009 coincides with the accelerated increase of urbanization and shrimp farms as well as the necessary associated infrastructure through the Jaguaribe estuary. This strongly suggests a relationship (Table II) between them (Figure S3) and the weakening of protecting measures resulting from the 2012 amendments of the BFC. Along the semiarid NE coast of Brazil, shrimp farming and, to a lesser extent, salt ponds respond for only 1 to 3% of the direct conversion of mangroves, notwithstanding its “permanent protection” legal status (Vanin et al. 2024). These figures are by far smaller than the dieback rates observed in the Jaguaribe Estuary. In fact, recent syntheses on the impact of shrimp aquaculture on mangroves have highlighted also their indirect impacts in several sites at Neotropics and in most S-SE Asia (Guimarães et al. 2009, Marins et al. 2020, Lacerda et al. 2019, 2021, 2024). But in some countries in SE Asia and in Ecuador, in South America (Hamilton 2013, Ofori et al. 2023), shrimp farm was the major cause of mangrove deforestation.

The area of intensive shrimp aquaculture has increased exponentially after 2006 on the mangroves of semiarid Brazilian northeast, and their cumulative impact has been extensively studied (See Lacerda et al. 2021, for a review). In the Ceará State, after 2008, salt flat and mangrove areas decreased respectively 37.12% and 29.37%, but shrimp pond area increased 24.73%, being this commodity production often related to degradation of coastal habitats and the loss of crucial ecosystemic services furnished by them (Lima et al. 2025). Although rarely directly converting mangroves into ponds, with the exception of opening intake and effluent channels, access roads and harbors, the shrimp aquaculture has been directly associated with increasing eutrophication and siltation and was related to decreasing canopy health and primary productivity, as observed by the decreasing of the Normalized Difference Vegetation Index (NDVI), also previously observed in the region by Marins et al. 2020. The vegetation index (NDVI), which compares photosynthetic activity to canopy structure of forests, strongly associates mangrove degradation with increasing shrimp farm area in many regions (Alatorre et al. 2016). In the Jaguaribe Estuary, Marins et al. (2020) showed a 15% reduction in the NDVI between 2003 (NDVI = 0.78) and 2008 (NDVI = 0.65) followed by a faster 70% reduction between 2009 and 2017 (NDVI = 0.2), fitting with an increase in shrimp pond area from 1,388 ha in 2003 to 1,992 ha in 2009 to 2,189 ha in 2018. Also, the excess of nutrients from shrimp farm effluents have been observed to change the essential biogeochemistry of nutrients, such as phosphorus, directly affecting their bioavailability (Fauzi et al. 2014). It is of notice the occurrence of a time-lag interval between the onset of shrimp ponds area (starting in 2006) and mangrove dieback, only evidenced by 2015. Similar correspondence between decrease on mangrove area and growth of shrimp farms was found in Ecuador (Rebolledo Monsalve et al. 2024), where production of this commodity also altered the physicochemical nature of water and sediments and the stability of these latter, leading to erosion and reduced carbon accumulation. Also, shrimp aquaculture has been associated to impacts on macrobenthic assemblages (Ribeiro et al. 2016, Rebolledo Monsalve & Verduga Vergara 2023) which harbor key functional groups of organisms.

On the other hand, expansion of urbanization seems to directly produce mangroves degradation (Figure 3), in agreement with the correlation analysis. Urbanization moves soil and changes watercourses, including tidal creeks, with direct impacts on mangrove from excess siltation and hypersalinity, as shown in many areas worldwide (Botero & Salzwedel 1999, Lacerda et al. 2019). Coincidently, the acceleration of the decrease of mangrove extension, occur during the onset of one of the longest extended droughts in the region between 2011 and 2017 (Figure 3), which probably contributed to decreasing mangrove health. Although, even after the drought period, mangroves keep on losing extension.

The expansion of urbanized areas showed a strong negative and significant correlation (p < 0.01) with mangrove extension. The relatively small area where the major urban area interacts with mangroves along the estuary, and the highly significant correlation coefficient observed between urban expansion and shrimp aquaculture (Table II), strongly suggest that increasing infrastructure to support activities far from urban centers, such as the recent established wind power generation fields and the expansion of shrimp farms are responsible for this strong correlation between urbanized areas and mangrove dieback. Infrastructure construction, mostly building roads and bridges, affects local hydrology, changing water renewal and increasing salinity, as demonstrated in many mangrove areas worldwide (Branoff 2017, Lacerda et al. 2021). Both shrimp (Table I) and wind farms have witnessed a boom in the region from 2006 onwards. Wind power has increased from 0.08 GW.yr-1 in 2006 to 236 GW.yr-1 in 2021 (Rodrigues et al. 2021). These increases required extensive infrastructure constructions, as clearly observed in the images available in Figure S3.

Table II shows that local anthropogenic drivers are also accompanied by a decrease in extension of dune fields. Worldwide, dune displacement depends on the intensity and duration of the dry season (Maia et al. 2005), generally more affected by drivers associated with global changes, rather than local human activities and are being intensified and prolonged due to climate change (Marengo et al. 2018). This has resulted in invasion and burial of mangroves by mobile sand dunes (Lacerda 2018, Lacerda et al. 2024). In the estuarine basin of the Jaguaribe river, dune fields, similar to mangroves, show a significant negative correlation with shrimp farming and urbanization (Table II), suggesting that direct changes, such as planting vegetation to fixate dunes, opening of roads and establishing physical infrastructure, may be associated with decreasing dune field area shown by the images.

In many semiarid regions, mangroves have shown to be more vulnerable to El Niño conditions due to its effects on decreasing rainfall (Duke et al. 2022). Drier conditions increase hydric stress, reduce continental sediment and nutrient supply and strengthens ocean forcing, strongly affecting mangrove development (Lacerda et al. 2020, 2024, Echeverry Hernández et al. 2025). Indeed, the years 2015 and 2016 were marked by strong ENSO occurrence, with high temperatures and decreased rain levels. Some mangroves throughout the neighbor State of Rio Grande do Norte showed degradation and/or dieback in the past decade, and according to close-living folks the time of this degradation was precisely the period between 2015-2016, characterized by high temperature and low rainfall. However, since some of these stands are bordered by shrimp ponds and showed signals of defoliator insect infestation (A. Ferreira, personnal observation), degradation cannot be confidently attributed to only one factor, instead, likely a synergy between them.

There was no significant correlation between ONI and mangroves or salt flats extensions, differently of dune field extension, which, as expected, was significantly correlated with ONI (rs = 0.674; p < 0.01). This result agrees with observation in other regions of the semiarid coastlines of NE Brazil, where increasing dune field cover were associated with stronger El Niño events (Maia et al. 2005) and suggest that with increasing strength of El Niño, dune encroachment can be a significant threat to mangroves in the Brazilian semiarid, and mangroves from semiarid regions in general, as proposed by Lacerda et al. (2024). On the other hand, it clearly suggests that ONI is not associated with mangrove cover changes in the Jaguaribe region, as confirmed by the lack of association between mangrove extension and ONI (Figure 4), and this result strengthens the role played by local anthropogenic drivers on mangrove area change in the region.

It is interesting to note that, although not significant at p < 0.05, the correlation analysis showed a positive association between ONI and mangrove extension, suggesting an increase in mangrove area following the aridity provoked by El Niño. This may be associated with mangrove migration landward, as observed in different estuaries of the Brazilian northeast region, including the Jaguaribe estuary (Godoy et al. 2018). Mangroves move upriver as increasing aridity results in high salinity extending inland, up to the first dam in Itaiçaba, where mangroves are colonizing river margins, as narrow fringes of L. racemosa, as seen in Figure S4.

A PCA analysis of all variables of interest (Figure 5) clearly shows the interaction between different land uses and the degree of impact of the intensity of El Niño, as expressed by the ONI index. As suggested by the correlation coefficients, shrimp farm area and urban areas clearly positively covary while the three natural land cover classes negatively vary with the local anthropogenic drivers. The longer arrow representing shrimp farms, suggest aquaculture with the stronger influence on mangroves and salt flats. Shrimp pond area and salt flats are closer and with high negative contribution of the vertical axis (PC2). As for the El Niño influence, the moderately negative correlation with PC1 (-0.460) suggests that years with El Niño (higher ONI) tend to be associated with lower PC1 values, expansion of natural environments such as mangroves and sand dunes. The weak correlation with PC2 (–0.224) indicates a less clear pattern, perhaps with some indirect influence on salt pans, but not strong enough.

In two decades of studies, some forest fragments of former huge Avicennia trees as well as mixed mangrove stands died so extremely fast that the functional degradation that preceded the physical one must have progressed extremely quickly, despite functional degradation being in general a slower process (Lewis et al. 2016). Rapid dieback is often result of far-reaching processes that disrupt hydrological features (e.g. tidal coverage, sedimentation) and/or soil biogeochemical processes, specially that performed by a crucial microbiota from which depends on mangrove nutrition and hence growth (Duke et al. 2022, Holguín et al. 2001).

Other forest fragments along the estuary also show signals of ongoing functional degradation, with decreased NDVI, high sedimentation, trees with partially dead canopies and unusually high density of woodborer holes in the basal portions, and lack of key macrobenthic species, like the great mangrove crab, Ucides cordatus. Populations of this crab, significant shredder of tree litter (and hence a crucial contributor to detritus input) and with burrower/bioturbator habits that oxygenate the soil, have their populations in decline in the estuary. Other crab and fish species that were more common in past years show clear decline, as confirmed by local fishermen. Despite the state of functional groups (FGs) of key organisms to mangrove structure and functionality (Ferreira et al. 2024) were not systematically assessed, evidences suggest an ongoing process of functional degradation, likely via disruption of ecological functions of some FGs, varying in intensity among different fragments. Mangrove soils at Jaguaribe probably suffer with disruption of the biogeochemical mediator microbiota (a crucial FG), due to chemical conditions found in the soil and interstitial water (Costa et al. 2013). The probable connection among FGs disruption and mangrove dieback needs further studies, to fully assess this indirect driver of mangrove degradation from land use changes in the Jaguaribe estuary.

CONCLUSIONS

The ‘run for gold’ of exotic shrimp Penaeus (Litopenaeus) vannamei intensive breeding has left, through direct, but mostly indirect effects, deep and accumulated impacts in northeastern Brazil mangroves. Harsh conditions of semiarid climate and hydrological dynamics pose challenges to mangroves in northeast Brazil and other semiarid mangroves, but human activities, like aquaculture, accelerate degradation rate of an ecosystem increasingly important to face climate change rising forcings. We showed how production and outflow of shrimp commodity and infrastructure for access, production and energy transport can converge to mangrove degradation of the Jaguaribe estuary in NE Brazil, which still harbours the largest mangrove extension of semiarid coast. This suggests that these anthropogenic drivers should be addressed in conjunct. This increases the concern over small northeastern mangrove stands subjected to the same drivers, since forest erosion and fragmentation can have great deleterious effects when reached by climate change drivers of unpredictable forcings. At the same time, it calls attention to governmental agencies over unseen degradation effects on mangroves, and to weigh whether the production of a high aggregate value commodity is preferable to the dying of a mangrove forest that furnishes countless goods and ecosystem services.

Our conclusions on the effects of non-nature friendly practices and outcomes of shrimp farming in mangroves are likely to apply to areas with similar environmental settings, e.g., semiarid regions worldwide, and particularly in the Latin America and Caribbean region, and our findings can be considered to avoid further degradation and improve conservation and management of these forests, at least to a regional scale.

SUPPLEMENTARY MATERIAL

Figures S1-s4.

Figures S1-s4.

Acknowledgements

The authors thank the members of the Coastal Biogeochemistry Laboratory for their help in the work. This work was funded with resources from the INCT-TMCOcean - National Council for Scientific and Technological Development - CNPq Proc. No. 405.765/2022-3 and Ceará Foundation for Support to Scientific and Technological Development - FUNCAP Procs. Nos. ITR-0214-0010901.00/23 and UNI-0210-00030.01.00/23.

• Data availability
  Data will be made available upon reasonable request.

References

• ALATORRE LC ET AL. 2016. Temporal changes of NDVI for qualitative environmental assessment of mangroves: Shrimp farming impact on the health decline of the arid mangroves in the Gulf of California (1990-2010). J Arid Environ 125: 98-109. https://doi.org/101016/jjaridenv201510010.
• BOTERO L & SALZWEDEL H. 1999. Rehabilitation of the Cienaga Grande de Santa Marta, a mangrove-estuarine system in the Caribbean coast of Colombia. Ocean Coast Manag 42: 243-256. https://doi.org/101016/S0964-5691(98)00056-8.
• BRANOFF BL. 2017. Quantifying the influence of urban land use on mangrove biology and ecology: A meta-analysis. Glob Ecol Biogeogr 26: 1339-1356. https://doi.org/101111/geb12638.
• BUNTING P, ROSENQVIST A, HILARIDES L, LUCAS RM, THOMAS N, TADONO T, WORTHINGTON TA, SPALDING M, MURRAY NJ & REBELO L-M. 2022. Global Mangrove Extent Change 1996–2020: Global Mangrove Watch Version 3.0. Remote Sens 14: 3657. https://doi.org/10.3390/rs14153657.
  » https://doi.org/10.3390/rs14153657
• CHANDER G, MARKHAM BL & HELDER DL. 2009. Summary of current radiometric calibration coefficients for Landsat MSS, TM, ETM+, and EO-1 ALI sensors. Rem Sens Environ 113: 893-903. https://doi.org/101016/jrse200901007.
• CONGALTON RG. 1991. A review of assessing the accuracy of classifications of remotely sensed data. Remote Sens Environ 37: 35-46. https://doi.org/10.1016/0034-4257(91)90048-B.
  » https://doi.org/10.1016/0034-4257(91)90048-B
• CONGALTON RG & GREEN K. 2019. Assessing the accuracy of remotely sensed data: principles and practices. 3rd ed., Boca Raton: CRC Press, 201. https://doi.org/101201/9780429052729
  » https://doi.org/101201/9780429052729
• COSTA BGB, SOARES TM, TORRES RF & LACERDA LD. 2013. Mercury distribution in a mangrove tidal creek affected by intensive shrimp farming. Bull Environ Contam Toxicol 90: 537-541. https://doi.org/101007/s00128-012-0957-4.
• DIAS FJS, LACERDA LD, MARINS RV & DE PAULA FCF. 2011. Comparative analysis of rating curve and ADP estimates of instantaneous water discharge through estuaries in two contrasting Brazilian rivers. Hydrol Proces 25: 2188-2201 https://doi.org/101002/hyp7972
• DIAS FJS, MARINS RV & MAIA LP. 2009. Hydrology of a well-mixed estuary at the semi-arid NE Brazilian coast. Acta Limnol Brasil 21: 377-385. http://wwwrepositorioufcbr/handle/riufc/67880.
• DUKE NC, MACKENZI JR, CANNING AD, HUTLEY LB, BOURKE A J, KOVACS JM, CORMIER R, STABEN G, LYMBURNER L & AI E. 2022. ENSO-driven extreme oscillations in mean sea level destabilize critical shoreline mangroves - An emerging threat. PLOS Climate 1-23. https://doi.org/101371/journalpclm0000037
  » https://doi.org/101371/journalpclm0000037
• ECHEVERRY HERNANDEZ JP, MANCERA PINEDA JE & SÁNCHEZ NÚÑEZ DA. 2025. Migration of deltaic and open water mangrove forests in response to environmental factors. Front Mar Sci 12: 1569857. https://doi.org/103389/fmars20251569857.
• FAUZI A, SKIDMORE AK, HEITKÖNIG IMA, VAN GILS H & SCHLERF M. 2014. Eutrophication of mangroves linked to depletion of foliar and soil base cations. Environ Monit Assess 186: 8487-8498.
• FERREIRA AC, ASHTON EC, WARD RD, HENDY I & LACERDA LD. 2024. Mangrove biodiversity and conservation: setting key functional groups and risks of climate-induced functional disruption. Diversity 16: 423. https://doi.org/103390/d16070423.
• FERREIRA AC, FREIRE FAM, RODRIGUES JVM & BEZERRA LEA. 2022. Mangrove recovery in semiarid coast shows an increase in ecological processes from biotic and abiotic drivers in response to hydrological restoration. Wetlands 42: 1-19. https://doi.org/101007/s13157-022-01603-0.
• FERREIRA AC & LACERDA LD. 2016a. Degradation and conservation of Brazilian mangroves, status and perspective. Ocean Coast Manag 125: 38-46. https://doi.org/101016/jocecoaman201603011.
• FERREIRA AC & LACERDA LD. 2016b. Reply to “On the impact of the Brazilian Forrest Code on mangroves: A comment to Ferreira & Lacerda (2016) by Ronaldo Ruy Oliveira Filho et al. Ocean Coast Manag 132: 170-171. https://doi.org/101016/jocecoaman201608003.
• FUNCEME. 2025. Fundação Cearense de Meteorologia e Recursos Hídricos Portal Hidrológico do Ceará. Fortaleza. Available at: http://www.hidro.ce.gov.br/
  » http://www.hidro.ce.gov.br/
• GODOY MDP, MEIRELES AJA & LACERDA LD. 2018. Mangrove response to land use change in estuaries along the semiarid coast of Ceará, Brazil. J Coast Res 34: 524-533. https://doi.org/102112/JCOASTRES-D-16-001381.
• GUIMARÃES AS, TRAVASSOS P, SOUZA FILHO PW, GONÇALVES FD & COSTA F. 2009. Impact of aquaculture on mangrove areas in the northern Pernambuco Coast (Brazil) using remote sensing and geographic information system. Aquac Res 41(6): 828-838. https://doi.org/101111/j1365-2109200902360x.
• HAMILTON S. 2013. Assessing the role of commercial aquaculture in displacing mangrove forest. Bull Mar Sci 89(2): 585-601. https://doi.org/10.5343/bms.2012.1069.
  » https://doi.org/10.5343/bms.2012.1069
• HOLGUÍN G, VAZQUEZ P & BASHAN Y. 2001. The role of sediment microorganisms in the productivity, conservation, and rehabilitation of mangrove ecosystems: An overview. Biol Fertil Soils 33: 265-278. https://doi.org/101007/s003740000319.
• LACERDA LD. 2018. Burial of mangroves by mobile dunes: a climate change threat in semiarid coasts. ISME/GLOMIS Electr J 16(2): 6-10. http://wwwglomiscom/ej/pdf/EJ_16-2pdf.
• LACERDA LD, BORGES R & FERREIRA AC. 2019. Neotropical mangroves: Conservation and sustainable use in a scenario of global climate change. Aquatic Conserv Mar Freshwat Ecosys 29(8): 1347-1364. https://doi.org/10.1002/aqc3119.
  » https://doi.org/10.1002/aqc3119
• LACERDA LD, FERREIRA AC, BORGES R & WARD R. 2022. Mangroves of Brazil. In: Das SC, Thammineni P & Ashton EC (Eds), Mangroves: Biodiversity, Livelihoods and Conservation. Springer: Singapore, p. 521-563. https://doi.org/10.1007/978-981-19-0519-3_20.
• LACERDA LD, FERREIRA AC, BORGES R & WARD RD. 2024. Challenges to mangroves of the equatorial semiarid coast of Brazil in the Anthropocene. Cambridge Prisms: Coastal Futures 2: e17. https://doi.org/10.1017/cft202416.
  » https://doi.org/10.1017/cft202416
• LACERDA LD, MARINS RV & DIAS FJS. 2020. An Arctic Paradox: Response of fluvial Hg inputs and its bioavailability to global climate change in an extreme coastal environment. Front Earth Sci 8: 93. https://doi.org/10.3389/feart202000093.
  » https://doi.org/10.3389/feart202000093
• LACERDA LD, WARD RD, GODOY MDP, MEIRELES AJA, BORGES R & FERREIRA AC. 2021. 20-years cumulative impact from shrimp farming on mangroves of Northeast Brazil. Front Forest Glob Chang 4: 653096. https://doi.org/10.3389/ffgc2021653096.
  » https://doi.org/10.3389/ffgc2021653096
• LANDIS JR & KOCH GG. 1977. The measurement of observer agreement for categorical data. Biometrics 33(1): 159-174. https://doi.org/10.2307/2529310.
  » https://doi.org/10.2307/2529310
• LEWIS RR, MILBRANDT EC, BROWN B, KRAUSS KW, ROVAI AS, BEEVER JW & FLYNN LL. 2016. Stress in mangrove forests: Early detection and preemptive rehabilitation are essential for future successful worldwide mangrove forest management. Mar Pollut Bull 109: 764-771. https://doi.org/10.1016/jmarpolbul201603006.
  » https://doi.org/10.1016/jmarpolbul201603006
• LILLESAND T, KIEFER R & CHIPMAN J. 2015. Remote sensing and image interpretation. Wiley ISBN: 978-1-119-12846-5.
• LIMA FD, FERNANDES RTV, OLIVEIRA JF, PINTO ARM, FERNANDES RTV, ALMEIDA LF, COSTA DFS & NASCIMENTO L. 2025. Mapeamento das áreas de apicum disponíveis para carcinicultura no estado do Ceará, Nordeste, Brasil. Rev Brasil Geogr Fís 18: 3173. https://periodicos.ufpe.br/revistas/rbgfe.
• MAIA LP, FREIRE GSS & LACERDA LD. 2005. Accelerated dune migration and eolian transport during el Niño events along the NE Brazilian Coast. J Coast Res 21: 1121-1126. https://doi.org/102112/03-702A1
  » https://doi.org/102112/03-702A1
• MAPBIOMAS. 2023. Projeto MapBioma: Coleção 80 da Série Anual de Mapas de Cobertura e Uso da Terra do Brasil 2023. MMA, Brasília. https://mapbiomas.org
  » https://mapbiomas.org
• MARENGO JA, ALVES LM, ALVALA RCS, CUNHA AP, BRITO S & MORAES OLL. 2018. Climatic characteristics of the 2010-2016 droughts in the semiarid Northeast Brazil region. An Acad Bras Cienc 90: 1973-1985. https://doi.org/10.1590/0001-3765201720170206.
  » https://doi.org/10.1590/0001-3765201720170206
• MARINS RV, FREIRE GSS, MAIA LP, LIMA JPR & LACERDA LD. 2002. Impacts of land-based activities on the Ceará coast, NE Brazil. In: Lacerda LD, Kremer HH, Kjerfve B, Salomons W, Marshall-Crossland JI & Crossland JC (Eds), South American Basins: LOICZ Global Change Assessment and Synthesis of River Catchment – Coastal Sea Interaction and Human Dimensions LOICZ Reports & Studies No 21, p. 92-98.
• MARINS RV, LACERDA LD, ARAÚJO ICS, FONSECA LV & SILVA FATF. 2020. Phosphorus and suspended matter retention in mangroves affected by shrimp farm effluents in NE Brazil. An Acad Bras Cienc 92: e20200758. https://doi.org/10.1590/0001-3765202020200758.
  » https://doi.org/10.1590/0001-3765202020200758
• MARINS RV, PAULA FILHO FJ, ESCHRIQUE SA & LACERDA LD. 2011. Anthropogenic sources and distribution of phosphorus in sediments from in the Jaguaribe river estuary, NE Brazil. Brazil J Biol 71: 673-678. https://doi.org/101590/S1519-69842011000400011.
• MARINS RV, PAULA-FILHO FJ & ROCHA CAS. 2007. Geoquímica de fósforo como indicadora da qualidade ambiental e dos processos estuarinos do Rio Jaguaribe: costa nordeste oriental brasileira. Quim Nova 30: 1208-1214. https://doi.org/10.1590/S0100-40422007000500029.
  » https://doi.org/10.1590/S0100-40422007000500029
• MEIRELES AJA, CASSOLA RS, VANINI STE, QUEIROZ LS. 2007. Impactos ambientais decorrentes das atividades da carcinicultura ao longo do litoral cearense, nordeste do Brasil. Mercator 12: 83-106.
• NETELER M, BOWMAN MH, LANDA M & METZ M. 2012. GRASS GIS: a multi-purpose open source GIS. Environ Model Software 31: 124-130. https://doi.org/101016/jenvsoft201111014.
• NOAA. 2025. National Ocean and Atmosphere Agency Climate Variability: Oceanic Niño Index. Available at: https://www.climate.gov/news-features/understanding-climate/climate-variability-oceanic-nino-index
  » https://www.climate.gov/news-features/understanding-climate/climate-variability-oceanic-nino-index
• OFORI SA, KODIKARA SKA, JAYATISSA LP, SANDUNI K, MADARASINGHE SK, MAFAZIYA NIJAMDEEN TWGF & DAHDOUH-GUEBAS F. 2023. What is the ecological footprint of aquaculture after 5 decades of competition between mangrove conservation and shrimp farm development? Aquat Conser Mar Freshwat Ecosys 33: 15-28. https://doi.org/10.1002/aqc.3897.
  » https://doi.org/10.1002/aqc.3897
• OLIVEIRA-FILHO RR, ROVAI AS, MENGHINI RP, COELHO-JÚNIOR C, NOVELLI YS & CINTRON G. 2016. On the impact of the Brazilian Forest Code on mangroves: a comment to Ferreira and Lacerda (2016). Ocean Coast Manag 132: 36-37. https://doi.org/101016/jocecoaman201608002.
• POGSON M & SMITH DH. 2015. Effect of spatial data resolution on uncertainty. Environ Model Software (ENMS) 63: 87-96. https://doi.org/10.1016/jenvsoft201409021.
  » https://doi.org/10.1016/jenvsoft201409021
• QUEIROZ L, ROSSI S, MEIRELES J & COELHO C. 2013. Shrimp aquaculture in the federal state of Ceará, 1970 e 2012: Trends after mangrove forest privatization in Brazil. Ocean Coast Manag 73: 54-62. https://doi.org/10.1016/jocecoaman201211009.
  » https://doi.org/10.1016/jocecoaman201211009
• REBOLLEDO MONSALVE E, VERDUGA L & HURTADO FERREIRA F. 2024. Wetland ecosystem service preservation? Geochemical changes in systems with mangroves and shrimp farms in the northern Ecuadorian coast. Sustainability 16: 11083. https://doi.org/103390/su162411083.
• REBOLLEDO MONSALVE E & VERDUGA VERGARA L. 2023. Water and Sediment Quality Changes in Mangrove Systems with Shrimp Farms in the Northern Ecuadorean Coast. Appl Sci 13: 7749. https://doi.org/103390/app13137749.
• RIBEIRO L, ECA G, BARROS F & HATJE V. 2016. Impacts of shrimp farming cultivation cycles on macrobenthic assemblages and chemistry of sediments. Environ Pollut 211: 307-315. https://doi.org/10.1016/j.envpol.2015.12.031.
  » https://doi.org/10.1016/j.envpol.2015.12.031
• RODRIGUES MVS, SILVA LMX, LIMA SHD & PEREIRA RF. 2021. Panorama do setor eólico no Ceará/Brasil - situação atual e perspectivas futuras. Intern J Devel Res 11: 47006-47010. https://doi.org/1037118/ijdr21932052021.
• SCHMIDT AJ, BEMVENUT CE & DIELE K. 2013. Sobre a definição da zona de apicum e sua importância ecológica para populações de caranguejo uçá Ucides cordatus (Linnaeus, 1763). Bol Téc Cient CEPENE 19: 9-25. https://www.icmbio.gov.br/cepene/images/stories/publicacoes/btc/vol19/btc-vol19pdf.
• SILVA JAA ET AL. 2012. Código Florestal e a Ciência: Contribuições para o diálogo. Sociedade Brasileira para o Progresso da Ciência e Academia Brasileira de Ciências, São Paulo, 124 p. http://www.sbpcnet.org.br/site/publicacoes/outras-publicacoes/CodigoFlorestal__2aed.pdf
  » http://www.sbpcnet.org.br/site/publicacoes/outras-publicacoes/CodigoFlorestal__2aed.pdf
• VANIN GT, LACERDA ER & MORI GM. 2024. Drivers of mangrove area change and suppression in Brazil from 2000 to 2020. Conserv Biol 2024: e14426. https://doi.org/101111/cobi14426.

Edited by

• Handling editor
  Alexander Kellner

Data availability

Data will be made available upon reasonable request.

Publication Dates

• Publication in this collection
  10 Apr 2026
• Date of issue
  2026

History

• Received
  25 June 2025
• Accepted
  11 Oct 2025