Free and open-source software for Geographic Information System on coastal management: a study case of sea-level rise in southern Brazil

This work assesses sea-level rise using three different models created on Free and Open-Source Software for Geographic Information System (FOSS4GIS). Based on regional projections of Special Report on Climate Change and Oceans and Cryosphere (SROCC) of the Intergovernmental Panel on Climate Change (IPCC), the models were applied to a case of study on Rio Grande do Sul coast – Brazil under different sea-level rise scenarios by the end of this century. The End Point Rate for QGIS (EPR4Q), calculates a shoreline projection using End Point Rate method. The Uncertainty Bathtub Model (uBTM), analyses the sea-level rise impact by the uncertainty of sea-level projections and vertical error of the Digital Elevation/Terrain Model (DEM/DTM). The Bruun Rule for Google Earth Engine Model (BRGM) predicts the shoreline position with sea-level rise, using topographic and bathymetric data from Unmanned Aerial Vehicles (UAV) and Coastal Modelling System (SMC – Brazil), respectively. The results indicated a maximum shoreline retreat for 2100 of -502 m and -1727 m using EPR4Q and BRGM, correspondingly. The uBTM using the land-use of Mapbiomas showed a maximum of 44.57 km2 of urban area impacted by the sea-level flood. This research highlights the possibility of performing coastal management analysis in GIS environment using non-commercial software.


Introduction
It is undeniable that most coastal areas of the planet are currently impacted by Sea-Level Rise (SLR) [1,2]. The average rate of SLR observed by tide gauges in the 20th century was 1.7 mm.yr-1 [3], and the average altimeter level from 1993 to 2016 was 3.3 mm.yr-1 [4].
Moreover, [5] estimates that more than 600 million people live in low-altitude coastal areas, and the number may increase to more than 1 billion by 2060. Future sea-level rise as a result of global warming can seriously threat the human activities in these regions, and the consequences of the climate change will continue even if the emissions of greenhouse gases decrease significantly [6]. It is well known that the SLR will increase coastal erosion along worldwide coasts [7][8][9][10][11][12][13]. However, in contrast to this general trend, the morphodynamics and sediment supply responses (shoreline adjustment) to new sea-level conditions will vary significantly among different coastal stretches [14].
The last report of the Intergovernmental Panel on Climate Change (IPCC) included several scenarios of regional sea-level changes, which make it possible to get a more precise analysis of SLR impacts at a local scale [15]. Besides, with the increase of areas affected by SLR worldwide, it is important that free and efficient alternative Geographic Information System (GIS) models and tools would be developed for diagnosis, monitoring, and management of the coastal zone. These models can help stakeholders in different scenarios and contexts of coastal morphodynamic trends.

End Point Rate for QGIS (EPR4Q)
One of the key analyses in coastal management is the study of shoreline trends. This study is essential to understand the changes in behavior and predict the position of the shoreline in the near future. The End Point Rate for QGIS (EPR4Q) is a tool created for QGIS 3.4 that performs the End Point Rate (EPR) method [16]. The EPR is one of several techniques of shoreline trend analyses (i.e., Linear Regression Rate, Shoreline Change Envelope, Net Shoreline Movement, Weighted Linear Regression) [17]. The advantage of the EPR is to predict future positions with just two shorelines, being a powerful tool to get a first impression of the impacts of SLR in a specific study area.

Uncertainty Bathtub Model (uBTM)
The Uncertainty Bathtub Model is a method based on the Bathtub Model (BTM) and Terrset Sea-level impact, which reveals the probability of an area to be flooded, determining a specified elevation as a single value of water surface or bathtub [18]. In this case, the Uncertainty Bathtub Model is a BTM classified by the uncertainty of sea-level projections and the vertical error of the Digital Terrain Model (DTM) to predict the sea-level flood in the future [19]. Modifications of BTM as those proposed by [20] and [21] are examples of similar models for GIS environment.

Bruun Rule for Google Earth Engine Model (BRGM)
The Bruun Rule for Google Earth Engine Model is a tool created on this platform that performs the Bruun Rule equation [19]. The equation [22] uses the berm height and the depth of closure (DoC) as a physical variable to calculate the sea-level rise. The berm height is defined by a horizontal plateau on the beach face or backshore, and the depth of closure is an empirical measurement of the seaward limit of significant cross-shore sediment transport on sandy beaches [23]. Subsequently, [24] proposed a formulation using extreme wave conditions to obtain the value of DoC. The Bruun Rule is a controversial physical formula and requires precaution to be applied due to its simplicity. The equation does not contemplate the sediment supply, is two-dimensional and only considers one value of SLR [25]. Besides, the method allows to have a preliminary analysis, taking into account that sandy beach volume in a wave-dominated regime is mainly constant (e.g., Rio Grande do Sul -Brazil Coastal Plain (RSCP) beaches [26]) and shows 70% of similarity of the shoreline retreat predictions in laboratory-scale experiments [27].
This paper aims to analyze the shoreline position in 2100 as well as to evaluate the areas of littoral flood in the Rio Grande do Sul coastal plain, southern Brazil. The analysis is based on the application of three informatic tools considering two of the scenarios included in the last IPCC report: Representative Concentration Pathway (RCP) 4.5, and RCP 8.5. It is intended that the application of the models can provide different perspectives for analyzing the impact of sea level rise and support future decision making in the region under analysis. major coastal lagoons (Patos, Mangueira and Mirim) ( Figure 1). The State has 5.4% of the Brazilian population with a total of about three million people [28] and occupies the third position of the Gross Domestic Product (GPD) of all federative units [29]. Regarding the natural and cultural heritage, the Rio Grande do Sul coastal plain (RSCP) has important natural park -Taim, Lagoa do Peixe, Itapuã, Camaquã, Mato Grande and Delta of Jacuíwith a very important biodiversity [30] and being a crucial area for migration birds of all around the world [31]. The RSCP has four indigenous villages and several archaeological sites of firsts human occupations in the region, with more than five thousand years [32,33]. From an economic perspective, beyond the important harbor activities, agriculture, industry, and tourism [29], the mineral extractions are growing in RSCP and related to mineral resources as titanium, iron, copper, peat, coal among others [34] (Figure 1).

Geological Setting
The RSCP has the most uniform coastal plain of the country and also the widest extensive sandy barrier-lagoon system of South America [38] that includes the Patos Lagoon, which is considered one of the largest coastal lagoon in the world [39]. The lagoon has a length of about 250 km, a maximum width of 40 km, a surface area near 10200 km², an average depth of 5 m and a drainage basin area greater than 200000 km² [40,41]. The system is connected to the Atlantic Ocean through a narrow channel, which is less than 1 km wide and about 18 m mean deep. Nowadays, this channel is protected by two jetties (in the northeastern Cassino beach zone, and in the southwest São José do Norte) with a length of 1 km each ( Figure 1). The sandy barrier system, with approximately 33000 km² and an average width of about 100 km, corresponds to almost 9.5% of the whole RS area. Since the beginning of the Holocene, the barrier system shows the coexistence of different stratigraphical framework (stationary, regressive and transgressive tendencies) [38] in response to sea level fluctuations, sediment supply and coastal processes. Each type of barrier seems to have had different shoreline displacement trends in the long term (>10³ years) [41].
Nowadays, the beach system associated to the sandy barrier receives minimal sediment input from the continent. The beaches are constituted by well-sorted fine quartz sand and show dissipative to intermediate morphodynamic states [42]. The width is very variable, which in some cases can compromise the ability to moderate the impacts of adverse conditions, such as storms. Most of the sediment load carried by rivers that drain to the coast is trapped in the adjacent lagoons and associated environments [43]. The Patos Lagoon estuary is the largest exporter of fine-grain sediments to the inner and outer shelf regions [44].

Oceanographic Setting
The coast of RS is a typical wave-dominated environment with a combination of high and moderate wave energy and a low micro tide regime [38]. In the past, these characteristics, coupled with the abundance of sediment and the gentle slope of the continental shelf, have established perfect conditions for coastal barriers along the coast [38].
The tide is semidiurnal, with an average range of about 0.5 m. Storm surges are frequent and can raise the water level by more than 1 m [45]. Based on a 30-year (1979 and 2008) deep-water wave modeling with the NOAA WAVEWATCH III®, the average wave directions are from south quadrant (~51%, the sum of annual occurrence), followed in order of importance by east quadrant (~43%, the sum of annual occurrence). South swell waves are more frequent in the autumn and winter seasons, while East quadrant waves are dominant in summer (ENE seas) and spring seasons (ENE swells). The waves exhibit an average significant wave height between 1 and 1.50 m, with a peak wave period between 6 and 14 s [46]. Along the wave breaker line of the RS coast, there is a variation of more than 60% in the significant wave height [38]. Waves generated by powerful storms (extreme events) can reach up to 5 m in height. Such high wave energy conditions occur on average every 1-2 years in the RS [47,48]. However, most storms (>80%) on the RS littoral generate waves with 2 to 3 m height [49]. The net longshore sediment transport is in the NE direction [50][51][52]. Based on the CERC (Coastal Engineering Research Center) formula, the net northward longshore transport is estimated from 0.60 to 2.93 Mm³.yr -1 along the RS coast [52,53].

Climatic Setting
According to Köppen climate classification system [54], the RS coast has a humid subtropical climate (Cfa), with hot summers (with an average of 21 to 24ºC) and cold winters (with an average of 12 to 14ºC), and well-distributed rainfall throughout the year (~1350 mm.year -1 ) [55,56]. The weather is mainly influenced by climate systems such as the South Atlantic Anticyclone (SAA), which produces the hot and humid air mass (Atlantic Tropical Mass -aTm) during the spring and summer months, and the Polar Migratory Anticyclone (PMA), which is responsible for the Atlantic Polar Mass (aPm), more frequent during the autumn and winter months [43]. The NE wind direction is predominant during the spring and summer. The W-SW wind direction is the second most frequent and is more constant in autumn and winter [57].

Materials and Methods
The analysis of the shoreline behavior in a sea-level rise scenario was performed using the Uncertainty Model for the oceanic and lagoon shorelines (Patos lagoon), while the EPR and the Bruun Rule methods were used only for the oceanic shoreline. Major cities were chosen for further analysis.
The predicted values of sea-level rise for the region for the period 2081-2100 according to the Special Report on the Ocean and Cryosphere in a Changing Climate (SROCC) [56] are: 0.46 m (0.31 -0.63) under Representative Concentration Pathway (RCP) 4.5 and 0.68 m (0.50 -0.90) under RCP 8.5. These values include Antarctica ice melt, glacio-isostatic adjustment (GIA), the gravitational and rotational effects, as well as the obtained results by [57] for glaciers, land water storage (LWS), and Greenland.

End Point Rate (EPR4Q)
Since the EPR only considers two shorelines in the prediction calculation, it was developed a method to select the best years from the period 1984-2016 of the differences of tidal flats data from [58]. The raster of tidal flats is a product that resulted from a global extent machine learning process of 700000 satellite images, which mapped the change in tidal flats over 33 years (1984-2016) [58]. The tidal flats data showed adequate to calculate the shoreline variation since the geological and oceanographic characteristics of Rio Grande do Sul coast, which shows sandy beaches and micro tides, allows quantifying the effect of sea-level rise using the tidal flats data.
The method created to apply in the tidal flats data allows identifying the years that showed an anomalous exchange of sediments that can influence the shoreline position. Also, the years with fewer changes in shorelines are adequate to be used in EPR. The tidal flats data were subtracted, and the difference was transformed into vector points. A heatmap was created to identify clusters with considerable changes and they were manually classified in a GIS environment (Table 1).

Uncertainty Bathtub Model (uBTM)
The model was created on Google Earth Engine (GEE), a cloud-based platform that offers high-performance computing resources for processing geospatial datasets [59]. The model considers the uncertainly of the vertical accuracy of the DEM and the SLR projections [19]. The final output is a terrain image smoothed with a circle kernel filter classified from 0 to 100 percent, where places with pixels values of zero assumed by the model present zero percent of possibility to be impacted by the sea-level flood. The places with a hundred percent of possibility to be flooded have a pixel value of 100. Therefore, the model was applied in a Digital Terrain Model (DTM) from Digital Elevation Model (DEM) ALOS World 3D-30m (AW3D30) of Japan Aerospace Exploration Agency (JAXA) mission [59], with 30 m of spatial resolution and vertical accuracy of 4.1 m [60]. The product was transformed to DTM using the modules for digital terrain model of QGIS software: filter tool and close gaps with stepwise resampling tool configured with the mean slope in the percentage of AW3D30 DEM that resulted on 1.40%, calculated with the statistics for raster tool.
The methodology to evaluate the impact of sea-level flood using uBTM consists of extracting the urban infrastructure class of Land Use MapBiomas [37], transformed into polygons with raster to polygons tool, and applying the union tool with de UBTM result of RCP 4.5 and RCP 8.5. Subsequently, the areas were calculated in square meters and hectares that allow predicting the sea-level flood impact by the end of the century. The urban infrastructure class of Mapbiomas was manually segmented in QGIS 3.4 into twelve groups according to proximity, context, and area ( Table 2) to improve the evaluation of the impacted area.

Bruun Rule for GEE Model (BRGM)
The application of Bruun Rule was performed using the data extracted from field surveys with the Unmanned Aerial Vehicles (UAV) Phantom 4 Pro. The photogrammetric surveys performed in twelve beaches (Curumim, Albatroz, Tramandaí, Cidreira, Dunas Altas, Bujuru, Cassino, Saritas, Verga, Concheiros, Hermenegildo and Barra do Chuí) (Figure 2) show 2.5 cm of spatial resolution and precision of 1 mm extracted with Global Navigation Satellite System (GNSS) -Real-time kinematic (RTK) System South Galaxy Model G1.
Each UAV survey site is located in a different coastal barrier geological setting according to the [61]'s classification that recognized progradational or transgressive barriers (the dune field migrates towards the mainland), agradational barriers (stationary tendency of dune field) and retrogradational or regressive barriers (the dune field migrates towards the ocean) ( Table 3). The Bruun Rule for Google Earth Engine Model (BRGM) allowed extracting the result of Bruun Rule Equation with a simple vector line and a raster with the topo-bathymetric information to obtain the berm height or the dune foot, and the depth of closure of active beach profile.
The berm height was determined manually on BRGM analyzing the slope variation of the DEM derived from UAV survey data ( Figure 2). The depth of closure was based on [65]'s results, using the same technique of morphological indicators, recognizing the main changes in slope to delimit the lower shoreface for the Bruun Rule. The results of the Hallermeier DoC equation [22] to the region are between 26.8 m and 37.9 m [62]. The bathymetry raster for the region was obtained exporting the data of the Coastal Modeling System (SMC -Brazil [63]) and interpolated on QGIS with Inverse Distance Weight (IDW) method. The berm height and DoC were extracted on the twelve profiles under RCP scenarios: 4.5 and 8.5 for 2100.

End Point Rate for QGIS (EPR4Q)
The visual analysis of the classification of the shoreline trend based on changes in tidal flats demonstrates that the best years to perform the EPR are 1992 and 2016 ( Figure  3). The shoreline displacement shows considerable variability during the next decades across the RS coast ( Figure 4).  Figure 4) that revealed accretion, the results have shown a mean shoreline retreat rate with values of -0.5 m.yr -1 in the south and middle of the RS coast, which seems to concentrate in the southern portion of the large-scale coastal projection. In the north, the shoreline is more stable with relative small retreat rates, and an accretion rate about 8.7 m.yr -1 can be pointed at the mouth of Patos Lagoon.  On the north coast, Capão da Canoa and Xangri-lá beaches presented higher shoreline retreatment (-245 m in total) ( Figure 5 -3 and 4). However, Torres, Arroio do Sal, Imbé, Tramandaí, Cidreira, and Balneário do Pinhal showed stable and positive values of the shoreline (maximum of 574 m) ( Figure 5). In the center of the map, the Balneário Mostardense ( Figure 5 -8) might be the most impacted place showing a shoreline with a maximum retreat of -449 m. This retreat can affect directly the city by the end of the century, due to the fact that nowadays, the distance from the coastline to the city is about 100 m.

Uncertainty Bathtub Model (uBTM)
3.2.1. uBTM with AW3D30 DTM -RCP 4.5 The results of the Uncertainty Bathtub Model, using the AW3D30 under the RCP 4.5, scenario showed a total of 3811 hectares (ha) of urban area with ≥13% probability of all RSCP being affected by the sea-level rise flood in 2100 ( Figure 6) ( Table 4).
The urban areas of Rio Grande group are the region with the highest potential to be exposed to sea-level rise impacts, showing 1095.11 ha with a probability ≥13%. It is possible to identify zones of potential flooding around the city, on the Port of Rio Grande and Cassino beach (Figure 6 -6).
The Littoral North Cities Group 2 (LNGC2) are the urban areas more impacted by sea-level flood with 8 ha showing 100% of probability and the second group in the list with a high total area of ≥13% of probability to be flooded, 969.14 ha. It is possible to observe a large urban area around the Tramandaí River that can be seriously affected (Figure 6 -2).
The Littoral North Cities Group 1 (LNGC1) is also highly exposed to sea-level flood impact showing 290.93 ha with a probability ≥13% (Figure 6 -1). In Pelotas area is possible to distinguish three important spots of future sea-level flood impact: the center of the city, Laranjal beach and Colônia-Z3, a traditional fisher's community (Figure 6 -12). The total area with a probably ≥13% to be affected with sea-level rise is 323.68 ha.
In the Metropolitan Area of Porto Alegre (MAPOA), all the critical locations are situated on the margin of Guaíba River, from Ipanema beach, in the south, to Guaíba city in the north region (Figure 6 -8). The total area with probability ≥13% of being impacted reaches 813.63 ha, and the urban zone is the third-largest potential stricken area.
The groups of urban areas of SJN (120.65 ha), Arambaré and SRS (65.19 ha), LNGC 3 (51.70 ha), BR and Tapes (35.11 ha), and SLS (33.63 ha) are also most likely urban area impacted in 2100, in contrast with Chuí and Hermenegildo (6.45 ha), and Mostardas (6.20 ha) that are the group with less than 10 ha of total area effect ( Figure 6). Considering the RCP 8.5 scenario, there is an increase of 64627 ha of zones with a probability ≥13% to be affected by the sea-level rise in 2100 (Figure 7).
The critical urban area of the Rio Grande group shows an increase of 153.96 ha flooded when compared with the previous scenario ( Table 5). The Metropolitan Area of Porto Alegre (MAPOA) is the one that presents the highest probability to be flooding by  (Figure 7 -5, 10, 3, 11 and 9, respectively) are also considered at risk, in contrast with Mostardas (9.07 ha) (Figure 7 -4) and Chuí and Hermenegildo (7.65 ha) (Figure 7 -7) where less than 10 ha are affected (Table 5). The Littoral North Cities Group 2 (LNCG 2) has a probability of ≥13% to be flooded that correspond to an area of 150.98 ha (Figure 7 -2). The LNCG 1 shows a growth of 53.27 ha (Figure 7 -1), and Pelotas of 65.07 ha (Figure 7 -12).

Bruun Rule for GEE model (BRGM)
The total median retreat obtained by the BRGM for the RSCP coast it's about -593.29 m (RCP 4.5 scenario) and -860.07 m (RCP 8.5 scenario). The results allow distinguishing two distinctive sectors ( Table 6). The first one (Profiles 1 to 6, Figure 8  The southern region of the RSCP is the most hazardous area (Profiles 7 to 12, Figure  8

Discussion
The application of the models gives three different perspectives of sea-level impact analysis for supporting decision-makers of Rio Grande do Sul coastal plain. EPR4Q results reveal that approximately 28.60% of the RSCP shoreline might show coastal erosion and 1.11% severe erosion (i.e. shoreline retreat >-0.5 m.yr -1 and >-3 m.yr -1 , respectively as stated in [12] 's classification). According to [61], from 2002 to 2016, about 4% of the RSCP shorelines showed values (>-2.73 m.yr -1 ) that is considered as severe coastal erosion. Therefore, the present data shows that results are lower in EPR4Q (1.11%). The difference can be explained by the fact that the value proposed by [61] is estimated by extrapolation of punctual measures of shoreline across RS coast, and the value in this study was obtained from a unique and integrated dataset, that analyzed 123532 transects of the entire RS coast, suggesting that the value 1.11% for severe coastal erosion is more precise.
The results of shoreline trends using EPR4Q show that the most affected cities are Balneário Mostardense, Hermenegildo, Barra do Chuí, Capão da Canoa, and Xangri-lá.
Balneário Mostardense displays a seasonal variance on shoreline movement related with the geomorphological characteristics that induce a slightly coastal inflection, and a greater exposition to the waves in storm regime [62]  The mean value does not differ much from [63], otherwise is more moderate, but is quite distinct from the -13.8 m found by [62] in the short-term analysis. The reason is due to the variability in shoreline displacement that occurs on temporal scales which range from months to years [65]. The method to analyze spatial variability through density maps avoids this problem because it visually identifies the variation of accretion and erosion between the years, which allows to select those with less variation caused by seasonal variability and extreme events. Moreover, the extracted shorelines between 1996 and 2001 shows similar distribution values as the ones found by [66] in the period 1997-1999, increasing the consistency of the method used.
A shoreline accretion trend was previously found in Torres (>3 m.yr-1 ), Tramandaí and Imbé (<1.0 m.yr-1 ), and Cassino (>~3 m.yr-1 ) using end point rate method [67]. The study performed by [68] in Hermenegildo, using the least square method, revealed a shoreline total retreat of -151.57 m in 2100 which is a value comparable with the mean -136. 22 and median retreat -164.68 m, calculated for the Hermenegildo in this work with EPR.
The uBTM analysis makes it clear that Rio Grande, MAPOA, LNCG 2, and Pelotas are the groups of cities with more probability (between 13% and 50%) to be impacted with sea-level flood under RCP 4.5, and RCP 8.5 scenarios by the end of the century. Additionally, LNCG 2, MAPOA, LNCG 1, and Rio Grande are the group of cities with the greater area (100% of probability) to be impacted by the sea-level rise flood under both scenarios due to geomorphological aspects, low altitude coast and great topographic uniformity. Nowadays, Tramandaí (LNCG 2) and Cassino (LNCG 1) are affected by washover episodes, storms surge induced by cyclones, and even meteorological giant waves as on 9 February 2014 at Cassino Beach, near the Rio Grande city [69][70][71]. The [72]'s study using a bathtub to evaluate the sea-level flood with high precision digital elevation model revealed the sea rise can impact the urban infrastructures by the end of the century as uBTM results pointed in both RCP 4.5 and RCP 8.5 scenarios.
The results given by the BRGM exceeded the retreat predicted in Hermenegildo in comparison to those obtained with the Random Shoreface Translation Model (RanSTM) [73][74][75][76]. [73] obtained a mean retreat distance of -521 m for 2100 using sea-level projections provided by the IPCC 2007-AR4. [77] found a mean retreat of -567 m using the projections of IPCC 2014-AR5 while [75] with the same sea-level projections released a mean retreat distance of -716 m due to take into account the effect of coastal hard protection structures. BRGM using regional projection of IPCC SROCC points to the 2100 horizon values of -1.18 Km (RCP 4.5) and -1.73 Km (RCP 8.5). The differences can be explained by the fact that sea-level trend values are not the same, and the RanSTM model is a stochastic model that can deal with uncertainties in the input data, being more precise than deterministic models, such as the Bruun model that ignored this aspect. The model also considers that the volume of water that drains the lagoon does not change as well as the tidal channel remains active, and assumes that the sea-level affects in Patos Lagoon are similar to those observed along the coast. In all models (EPR4Q, uBTM, and BRGM) it is also assumed that the existing coastal defenses (i.e. jetties) and other anthropogenic interventions in southern region (i.e. dredging and nourishments operations) are kept by 2100. Locations as Cassino beach, Imbé and Tramandaí are particularly vulnerable to changes in sediment input, interventions in the jetties (extension or reconfiguration) and dredging operations because they can change the sediment supply to the beach system. The impact of other extreme events can be more aggressive in the cities when analyzed the UBTM results. It is essential to prevent the urban expansion in the following years in areas near the sea and margins of Patos Lagoon, in order to avoid an increase in exposure to the hazard processes and the risk level.
There are other important issues that were not purposes of this preliminary study. These aspects are related to possible changes in climate and oceanographic settings, such as the mean direction of the wind, the angle of incidence of the waves, longshore currents direction or storm regime, which must also be considered in future coastal planning.

Conclusions
The use of Free and Open-Source Software for GIS (FOSS4GIS), considering the new IPCC information about the SLR for the RS coast, allowed to perform a complete and unprecedented analysis in all coastal plain extension. The use of one elevation model, two shoreline changes analysis, and three different sea-level models brought robustness to the results.
It can be stated that using the EPR4Q model is predicted a maximum shoreline advance of 795 m in some areas and a maximum shoreline retreat of -502 m by the end of the century. The shoreline retreat can majority threatening the cities of Balneário Mostardense, Hermenegildo, Barra do Chuí, Capão da Canoa, and Xangri-lá.
The sea-level flood analysis using the UBTM estimates that the urban areas can lose 38.11 Km 2 and 44.57 Km 2 in 2100 under the RCP 4.5 and RCP 8.5 scenarios, respectively. Examples of cities with high potential to be impacted by the sea-level flood are Tramandaí, Imbé, Porto Alegre, Pelotas, and Rio Grande.
The BRGM shows that the southern part of RSCP, where the cities of Cassino, Barra do Chuí, and Hermenegildo are situated, can be more affected by sea-level rise. The minimum retreat found was -298.46 m (RCP 4.5), and the maximum retreat noticed was -1.7 km (RCP 8.5) for the end of 2100.
Other studies have already identified coastal hazards as coastal erosion and flooding in the same cities found in this work, proving that is possible to create a completed sea-level analysis using only free and open-source software. In addition, it is prudent for local coastal managers of other cities in RSCP that exhibit some value of impact to pay attention to the increase of coastal exposure.