Dynamics of mangrove forest distribution changes in Iran

Natural resources and ecosystems provide various functions and services for human society in different regions (Falahatkar et al. 2017; Buonocore et al. 2020). Human activities, climate change, and the reduction of a large number of natural resources are the most important causes of the destruction and degradation of these resources (Taleshi et al. 2019; Mousavi & Falahatkar 2020; Mousavi et al. 2022). Marine and coastal ecosystems include a wide range of goods and services that provide vital food supplies for millions of people in the world (Haines-Young & Potschin 2018). Mangrove forests, as one of the highly productive coastal ecosystems and an important tropical environment, provide a large variety of goods and services, including being the source of wood products for many coastal communities, essential habitats for marine fishes, high rates of carbon sequestration in soil, high intrinsic natural productivity, valuable natural resource with distinctive diversity, and coastal defense service (Harris et al. 2017; Nazarnia et al. 2020; Kamil et al. 2021). Unfortunately, these forests have been threatened due to the development of aquaculture, agriculture, urban development, oil exploration activities, and industry worldwide and have lost significant amounts of their areas in recent decades. (Abbaspour et al. 2011; Haddad et al. 2015; Mokhtari et al. 2015; Kourosh Niya et al. 2019a). In addition to human activities, climate change is one of the most important causes of the change and destruction of mangroves (Mousavi et al. 2017; Cinco-Castro & Herrera-Silveira 2020; Banerjee et al. 2021).

Under long-term impacts of climate change, marine ecosystems are becoming environments with new conditions that may affect current conservation mechanisms (Johnson & Watson 2021). In the future, mangrove forests are also expected to be significantly affected by climate change (Ward et al. 2016). Due to the rising sea level as the most critical factor in climate change, the distribution and diversity of mangrove communities could change and consequently lead to fewer mangrove carbon stocks (Adame et al. 2021; Davar et al. 2021). Climate change is also predicted to have diverse effects on different mangrove communities and some regions are likely to experience more adverse effects (Ward et al. 2016). Changes in the distribution, health and productivity of mangrove forests can largely affect the presence and abundance of other species, as well as the ecosystem services of the forests (Brandt et al. 2019; Garcia et al. 2020; Wang & Gu 2021). Nowadays, a significant amount of research uses the species distribution modelling method to identify the effects of climate change on species distribution and to identify conservation priorities (e.g., Amiri et al. 2021; Ebrahimi et al. 2021; Ghane-Ameleh et al. 2021; Sales et al. 2021; Tehrani et al. 2021) and it is even used in other fields of science as well (Allahbakhshi et al. 2020; Firozjaei et al. 2021). In this regard, species distribution models (SDMs) have been ranked as one of five research methods in biological sciences (Warton et al. 2013) that utilize the presence/absence of data to predict species distribution and thus cover an important part of ecological studies (Warton et al. 2013; Mohammadi et al. 2019; Ilanloo et al. 2020). Therefore, to recognize priority areas for mangrove conservation actions, SDM results can be used under current and future climate conditions (Rodríguez-Medina et al. 2020; Lemes et al. 2022).

In this study, the SDM tool was used to assess the potential impacts of climate change on the distribution of mangrove forests along the southern coasts of Iran. Although studies have been conducted on mangroves in a range of subjects in the past, a complete picture of the entire region of mangrove distribution is not available. (Ghayoumi et al. 2019; Forouzannia & Chamani 2022). This study aims to investigate the current distribution of the Avicennia marina (dominant species in mangroves across habitats in Iran) and evaluate the consistency and variations in the forecast potential distributions of this species in Iran under the IPCC's future climate scenario (2070: Representative Concentration Pathway (RCP) 8.5, CCSM4). The results will help us understand how mangrove forests will experience the effects of climate change in concerning distributional shifts.

The habitat of A. marina is located on the shores of the Persian Gulf and Oman Sea. The natural mangrove habitats in Iran range from areas spanning the latitudes of 24° to 35 °N in latitude and 51° to 58 °E in longitude (Figure 1). These forests in Iran are categorized into the ‘non-industrial’ and ‘conservation tree’ communities (Safa Eisini 2006; Tiab et al. 2014). The distribution of natural mangrove forests in Iran starts in the eastern part of the Oman Sea at the Iranian border (Guater/Gwadar Bay) and ends in the west of the Persian Gulf and the Bushehr Province. Generally, mangrove forests are distributed in southern Iran on the coasts of the Persian Gulf and Oman Sea and in three provinces (from east to west): Sistan and Balouchestan, Hormozgan and Bushehr with 1,380, 19,381, and 325 hectare areas, respectively (Safa Eisini 2006). Khuzestan is the western-most province to the coast that does not have natural mangroves (Ghayoumi et al. 2019). A. marina forests in Iran are categorized as non-industrial and conservation tree communities; their exploitation only happens in the forms of harvesting of branches, beekeeping, aquaculture, recreational, and medicinal use. Parameters affecting the distribution of mangrove species include soil salinity, water penetration, physical and chemical properties of soil, water cycle, biological interactions, nutrients, and climate effects (Danehkar 1998; Safa Eisini 2006).

To construct the SDMs, we applied climatic, oceanographic, and environmental variables relevant to the species ecology (Supplementary material, Appendix 1). Climatic variables were downloaded from WorldClim (Fick & Hijmans 2017; https://www.worldclim.org/data/worldclim21.html). Environmental variables were acquired from the Iranian Ports and Maritime Organization and Oceanographic variables were downloaded from Auckland data (https://gmed.auckland.ac.nz/). To equalize the layers, ArcGIS 10.8 (30 arcsec resolution, ∼1 km²) was used to calculate the Euclidean distances of all pixels. During a year of field investigation, the intertidal zone in the Persian Gulf and the Sea of Oman was also studied, and each natural and planted polygon was recorded as a point. Finally, only natural points (32 points) were used as model inputs (Figure 1). The variance inflation factors (VIFs) in the niche-based model were calculated using the R programme (Naimi et al. 2014). This method allows us to avoid collinearity that could bias forecasts made by the niche-based model. A VIF larger than 10 indicates a collinearity problem in the model and by taking VIF>10 into account, the species’ linear variables were removed (Naimi et al. 2014). Thus, we retained eight variables that were applied throughout this study (see Supplementary material, Appendix 1).

The CCSM4 was used for future (2070) climatic conditions (Gent et al. 2011) since the CCSM4 climate model is widely used in Southwest Asia (e.g., Mohammadi et al. 2019; Ebrahimi et al. 2022; Ebrahimi & Ahmadzadeh 2022), and is one of the most efficient global climate projections to forecast the impact of future climatic changes on the distribution pattern of animal and plant species (e.g., Mohammadi et al. 2019). For the future climatic conditions, RCPs 8.5 climate scenarios based on Intergovernmental Panel on Climate Change (IPCC) reports were used from the WorldClim database (1.4; Fick & Hijmans 2017). The RCP 8.5 predicts the effects of climate change on pessimistic conditions, indeed RCP 8.5 depicts a rising radiative forcing pathway leading to 8.5 W m^–2 by 2070 (Shekede et al. 2018).

Species distributions were modelled using the Maximum Entropy (MaxEnt) approach, with presence-only species occurrence records and randomly generated background points (Phillips et al. 2006). Various SDMs have been used to evaluate the ecological requirements, ecological responses, and distribution areas (Sheykhi Ilanloo et al. 2021). Among these different modelling approaches, Maxent is widely used since it performs better (Pearson et al. 2007). Not requiring the points of absence is one of the main advantages of this approach (Felicísimo et al. 2013). This method estimates the most uniform distribution of sample points for the limitations of the data (Phillips et al. 2006). Also, some studies have shown that Maxent performs equally to an ensemble approach, but with less computational time, so an ensemble approach is not superior to this model (Hao et al. 2020; Kaky et al. 2020).

In this study, data were randomly divided into 70% records for model calibration and 30% for model evaluation (Naimi & Araújo 2016). To investigate the importance of variables, the Jackknife test was used. Furthermore, receiver operating characteristics (ROC) and the area under the curve (AUC) were used for the performance evaluation of the model (Elith et al. 2011; Naimi et al. 2014; Zhao et al. 2021). It is necessary to mention that the range of the AUC is 0.5–1.0 and the closer it is to 1 indicates the better performance of the model (Phillips et al. 2006).

The results of assessing MaxEnt model performance showed that AUC was equal to 0.907 for the training and 0.894 for the test data, which shows a pretty good prediction of the habitat distribution of A. marina in this study area. This model shows that suitable habitats for the A. marina in Iran are located on parts of the eastern coast of the Persian Gulf and the middle parts between the Persian Gulf and the Sea of Oman. Other habitats in the Persian Gulf are either completely unsuitable or have very low suitability for this species (Figure 2). We found that temperature annual range with a contribution of 21.4%, sea surface temperature range with a 16.4% contribution, and slope with a 15.71% contribution were the most important variables in determining the distribution of A. marina in our study area (Figure 5). According to the results, the highest probabilities of species occurring are in: slope <10%, sea surface temperature range <5.78, chlorophyll A (min)=1, BIO 19=6.5, BIO 12=9, BIO 07=18.5 and tidal range=0 (Figure 4).

This study of the effects of climate change on suitable habitats for A. marina in Iran using CCSM4 general circulation models showed that this species will lose a significant portion of its suitable habitats in the period of 2060–2080 under RCP 8.5 (Figure 3). The current potential suitable habitat of the species is 12.5% of the study area but will decrease to 2.1% of the study area. The highest loss of suitable habitats would be in the eastern parts of the Persian Gulf and Oman Sea (Figure 3).

The global mangrove area is declining due to increased human activities, climate change, and adverse environmental conditions (Kourosh Niya et al. 2019b; Goldberg et al. 2020). The mangrove ecosystem on the southern coasts of Iran is the last distribution edge range of these forests and has harsh and dry environmental conditions close to the species’ climatic tolerance limits (Schile et al. 2016; Mafi-Gholami et al. 2020). Therefore, they demonstrate lower productivity and biomass than mangrove communities in humid areas and are more sensitive to any additional pressure (Etemadi et al. 2018; Adame et al. 2021). Hence, climate change as extra stress can have irreversible impacts on these communities (Etemadi et al. 2021).

This study discusses the distribution of A. marina under climatic factors and oceanographic conditions. However, it should be considered that various factors are involved in mangrove distribution on a national and local scale (Jayanthi et al. 2018). The results showed that the most important variables in the distribution of A. marina are the temperature annual range, sea surface temperature range, and slope, which are consistent with previous studies (Servino et al. 2018). Mangrove forests are located in estuaries and the coasts where the environment is harsh and are highly resilient to changing conditions (Alongi 2015). However, one of the most important possible drivers of their extinction which has been identified in the distribution of A. marina is the increased thermal stress that causes an increase in evaporation and leads to more salinity and ultimately reduces tree growth and leaf gas exchange as well as seedling survival (Servino et al. 2018).

The slope is also identified as one of the effective natural factors influencing mangrove development. Light slope and tidal flat habitats are more suitable conditions for the distribution of the species (Figure 4) (Chen et al. 2017; Duncan et al. 2018).

The previous study has shown that the eastern part of the Persian Gulf, the western part of the Oman Sea, and a few spots in the easternmost part of the coastal areas in Hormozgan province are the most suitable areas in terms of ocean characteristics for mangrove development (Khayrandish et al. 2015) which is consistent to our findings in this study.

Geological evidence shows that mangroves have adapted to previous climate changes (Alongi 2015; Woodroffe et al. 2016; De Palma et al. 2017). However, some species are susceptible to changes in hydrological or tidal levels that can lead to damage or destruction (Blasco et al. 1996; Martínez-López et al. 2015; Cui et al. 2021).

In general, the results of this study depicted the losses of suitable habitat area for A. marina in Iran based on 2061–2080 future climate scenario (RCP 8.5) compared to the current condition and climate and oceanographic changes emerging as critical drivers of the future distribution causing biological range shifts toward the Oman Sea.

Alongi (2015) showed that mangrove forests in humid climate conditions experience less stress, less salinity, and more freshwater and therefore, their biomass and productivity are greater than the communities residing in more arid areas like the southern coasts of Iran.

Predictions for the Persian Gulf, which is considered a dry region, show that the area of mangrove forests will be declining due to the synergistic impacts of salinity and aridity, shortage of freshwater and frequent exposure to the critical temperature threshold (Alongi 2015; Ward et al. 2016).

According to the results, A. marina conservation is likely to be a difficulty in the future, and future temperature increase could be the most important factor in the species’ distribution (Figure 5). Temperature increase could limit the width of mangrove forests by salt marsh changes in a spatial context and, as a result, changes the species composition (Saintilan et al. 2014; Alongi 2015; Godoy & Lacerda 2015).

The future suitable habitat projections can help various aspects of decision-making in action management, resource management, conservation, restoration, and sustainable use. Since there is little information on the effects of climate change on mangrove forests in this area (Ward et al. 2016), this research provides valuable information for climate change mitigation initiatives and future adaptation of habitats, as well as the determination of protected areas and management areas to protect mangroves on the coasts of the Oman Sea and the Persian Gulf. In this study, only climatic data were used to determine future species distribution. Therefore, it is suggested that small-scale research be performed based on the appropriate combination of variables to provide more realistic outputs. On the other hand, it is suggested that in addition to ecological aspects, social and behavioural aspects should be considered in the management and protection of these forests (Garcia et al. 2020).

This research was funded by the Research Center for Environment and Sustainable Development, Department of Environment of Iran. We are grateful to Dr Farhad Hosseini Tayefeh for fieldwork assistance; and the Department of Environment of Hormozgan, Sistan and Baluchestan, and Bushehr provinces for their logistical support. We also thank the Iranian Ports and Maritime Organization for providing the data on the environmental layers.

All relevant data are included in the paper or its Supplementary Information.