European Geologist Journal 44

Multi-source data integration of Lower Cretaceous units with geothermal potential in Lisbon region, Portugal, to support geological modelling

by Ana Ramada1, Rayco Diaz2,3 and João Carvalho2

 Departamento de Ciências da Terra, FCT Universidade NOVA de Lisboa, Campus da Caparica, 2829-516 Caparica, Portugal;

2  Laboratório Nacional de Energia e Geologia (LNEG), Bairro do Zambujal, Alfragide, 2610-999, Amadora, Portugal;

3 Instituto Volcanológico de Canarias (INVOLCAN), 38400 Puerto de la Cruz, Tenerife, Spain.



The Lisbon region, the urban area with the highest population density and energy demand in Portugal, presents a favourable geological environment for geothermal purposes. The objective of this study was to integrate all the information known from different sources to support the construction of a geological model of the Lower Cretaceous units, which reveal the geothermal potential. The methodology developed in this study allows the integration of data from different sources into a single georeferenced and refined database. The eastern sector of the city of Lisbon, identified as the most promising area for geothermal applications, shows a lack of geological information at depth. However, a future junction of existing seismic profiles with the data obtained in the present study could resolve this issue.

  1. Introduction

It is widely known that large sedimentary basins are favourable for the existence of deep aquifers which, with a medium or even smaller geothermal gradient, are also susceptible to be exploited as low-enthalpy geothermal reservoirs. The Lisbon region, located in the western Meso-Cenozoic sedimentary basin (Rassmussen et al., 1998; Dias et al., 2013) presents deep aquifers in sedimentary formations, very favourable for geothermal purposes (Correia et al., 2002; Marrero-Diaz et al., 2015). Within the lithostratigraphic units identified in the Lisbon region, the geothermal potential of the Lower Cretaceous stands out with its relatively low mineralised water (salinity <1 g/l) and with a temperature of 50 °C at 1,500 m depth. These units were already being exploited in the 1990s for geothermal purposes in two concrete cases (Carvalho & Cardoso, 1994).

The main objective of the present study was to obtain a methodology to support the construction of three-dimensional geological models aiming to estimate representative surfaces of the lithostratigraphic units of the Early Cretaceous in order to infer their geometry in the Lisbon region. For this purpose, all known information from various sources – geology, hydrogeology and geophysics – was integrated with the intent of filling the lack of data in depth and identifying the most promising sites for geothermal purposes.

The region of Lisbon, with approximately 960 km2, is the urban area with the highest population density in Portugal (corresponding to about 19 % of the inhabitants of the country), which results in a strong demand for energy; in this way, geothermal energy could be seen as a possible environment-friendly energy source in this area.

  1. Study area

From a tectonic-sedimentary point of view, the Lisbon region is part of the southern sector of the Western Meso-Cenozoic border, with formations that belong to the Lusitanian Basin, partially overlain by Cenozoic sediments of the Baixo Tejo Basin (Figure 1). The Lusitanian Basin, deposited in a tectonic pit originated by the tilting of the Hercynian Massif, has been constituted by a >3 km thick normal sequence of sediments since the Triassic until today (Rasmussen et al., 1998). The Baixo Tejo Basin represents a normal sedimentary sequence with a thickness of about 1,500 m from the Paleogene until today (Dias et al., 2013).

Figure 1: Geological map of the study area, with the study area limit in red, obtained from the junction of Sheets 34A, 34B, 34C and 34D of the Geological Map of Portugal at 1:50,000 scale (LNEG, 1993, 2011, 1999, 2006). Coordinate system Lisboa Hayford Gauss IGeoE used here and in all maps below.

Due to the vast number of geological formations identified in the Lisbon region, only 8 lithostratigraphic units were chosen to simplify and group (Table 1).

Table 1: Lithostratigraphic units grouped in the study area and acronyms of the formations considered from Sheets 34A, 34B, 34C and 34D of the Geological Map of Portugal at 1:50,000 scale (LNEG, 1993, 2011, 1999, 2006).   

From all the lithostratigraphic units identified, Lower Cretaceous units, namely the Grés de Almargem and Barremian-Berriasian units (Table 1), have the greatest interest considering their geothermal potential (Marrero-Diaz et al., 2015). The Grés de Almargem lithostratigraphic unit is mainly constituted of ferruginous sandstones, pellets and conglomerates, interbedded with limestones and calcareous marls (Rey, 1993; LNEG, 2011). Generally, it is a sequence of sandstones, carbonates and sandstones again that has an average thickness in the order of 100 m, with relatively significant lateral variations. The Barremian-Berriasian unit groups several sedimentary formations composed essentially of limestones, marls, sandstones and pellets, with an average thickness of 200 m (LNEG, 1993).

According to Marrero-Diaz et al. (2015) and references therein, the Grés de Almargem unit behaves as a multi-layered semi-confined, often artesian, dual-porosity aquifer with coexisting intergranular and fracture circulation and effective porosity between 7 and 18 %, allowing exploration flow rates up to 50 l/s. The statistical study of 76 groundwater wells exploiting Grés de Almargem unit in the Lisbon region shows thicknesses between 60 m and 229 m with a mean value of 135 m, and transmissivities between 1 and 386 m2/d with a mean of 26 m2/d.

3. Multi-source data integration

3.1. Pre-processing

In an initial phase, a survey and selection of all the geological, hydrogeological and geophysical information available in the Lisbon region was carried out, from which the data summarised in Table 2 was selected.

3.2. Processing

The second phase consisted of the organisation and integration of previously selected data into a single georeferenced database using the ArcGIS ESRI platform, through the steps described below:

3.2.1. Geological data integration

The geological data integrated in the database belong to the geological mapping of Portugal at 1:50,000 scale (Table 2). This data was analysed, interpreted and treated using the ArcGIS ESRI platform and various automated ArcToolbox tools.

Table 2: Bibliographic summary of selected data.

Conversion of the polygons of lithostratigraphic units into points

The respective cartographic boundaries of the Grés de Almargem and Barremian-Berriasian units were unified and exported as polygons, which were later subdivided into points. Longitude (x) and latitude (y) in the Lisboa Hayford Gauss IGeoE system were assigned to each point thus obtained, and, finally, the elevation (z) was also assigned, based on the digital global elevation model (Aster GDEM) with approximately 30 m of spatial resolution.

Refinement of boundaries

In order to identify correspondence points with top and bottom of the lithostratigraphic units, a new field, designated as horizon, was added to the georeferenced database obtained from the previous steps. The coding assigned was 1 for the top, and 2 for the bottom.

There are points that do not correspond with the boundaries of top or bottom of the lithostratigraphic units; to solve this incongruence a refinement of the boundaries was manually performed through the ArcGIS ESRI platform, removing less representative points from the georeferenced database, for example, those coincident with contacts with volcanic extrusives or Holocene alluvial deposits. Figures 2 and 3 show Grés de Almargem and Barremian-Berriasian points before and after their refinement.


Figure 2: Grés de Almargem points before (left) and after (right) the refinement of boundaries.


Figure 3: Barremian-Berriasian points before (left) and after (right) the refinement of boundaries.

3.2.2. Hydrogeological data integration

For the present study, more than 500 reports of groundwater wells were consulted and analysed in the study area. After a critical review of each report, based on the geological and hydrogeological knowledge of the surrounding area, 251 wells were considered reliable and interesting for the present study (Figure 4).

Figure 4: Location of groundwater wells with information from the top and bottom of the various units in the study area.

A geological correspondence was proposed between the lithology crossed by the groundwater wells and the top and bottom of the lithostratigraphic units under study, as well as of the other units present in the region (Table 1). It should be noted that errors may exist in the proposed geological correspondence, due to the repetition of lithological sequences and the existence of asynchronous formations or lateral variations. Thus, the presented selection may be improved in the future with the introduction of new knowledge.

Analogously to the geological data integration described previously, codification (1 – top; 2 – bottom) was also assigned to each limit. Additionally, codes 0.5 and 1.5, corresponding to the top and base of the groundwater well, were assigned and denominated “Relative top” and “Minimum bottom”, respectively. These new horizons force the model to incorporate the information from the wells,, but it should be noted that they are not real boundaries.

3.2.3. Geophysical data integration

Multiple 2D seismic reflection profiles acquired for hydrocarbon prospection in the northern zone of the Lisbon region are available (Rasmussen et al., 1998; Carvalho et al., 2005). However, only 4 reflection seismic profiles are within the study area (Figure 5). These seismic profiles were acquired by Veritas in 1963 (AR-28 profile) and by the company Petróleos de Portugal (Petrogal) in 1981 (AR12-81, AR16-81 and AR17B-81 profiles) and initially interpreted by Walker (1983).

Figure 5: Location of seismic profiles in the study area.

Interpretation of the seismic profiles

The interpretation started with the seismic to well tie using well data in the study area and also using previously established seismic-stratigraphy packages identified by other authors in other seismic profiles outside the study area (Carvalho et al., 2005; 2011) that intersect the profiles used in this study. Afterwards, the interpretation proceeded by checking the intersections between each seismic profile and using geological surface information such as geological contacts, faults, etc. Available gravimetric and magnetic data (Carvalho et al., 2011) was also used in the interpretation, which allowed the identification of salt domes, igneous structures and the pre-Mesozoic basement.

Analysis of seismic profiles

Resolution of the seismic information is a crucial parameter for correct analysis of the seismic profiles. Considering an average propagation velocity of the P-waves of 2,500 m/s (corresponding to the homogenisation velocity identified in the considered profiles) and a frequency of 10 Hz, an average vertical resolution of 60 m was obtained.

Strong lithological subsurface contrasts are traduced in the seismic profiles by a strong reflector. However, chronological contacts do not always correspond to strong reflectors if the chronological boundary represents a continuous depositional episode and/or does not correspond to a sharp lithological boundary (Carvalho et al., 2005). To complement and calibrate this information, tops of the horizons of the Grés de Almargem and Barremian-Berriasian units (and of the other units present in the region) were also represented in the seismic profiles (Figure 6).  These horizon tops were obtained from intersecting seismic profiles previously reprocessed and reinterpreted by Carvalho et al. (2005; 2011) and from nearby groundwater wells (considering a buffer of 200 m between the wells and the seismic profiles).

Since the vertical scale of the seismic profiles corresponds to the traditional two-way travel time of the reflected waves, this last task involved the depth-time conversion of the groundwater well information. Due to the lack of check-shots and/or velocity logs in the wells that would allow the accurate knowledge of propagation velocities, the adopted criterion was to consider a homogenisation velocity of 2,500 m/s above the seismic datum (0 milliseconds – corresponding to approximately 150 m depth) and different velocities below 0 milliseconds obtained through the use of well data (Carvalho et al. 2005).

Next, a preliminary structural interpretation was performed, and several probable faults were identified, some of them with no surface correspondence in the geological maps of the study zone (Figure 6), though most of the major faults that were identified could be matched in the geological surface maps. Previously unknown salt domes and igneous structures were also identified.

Figure 6: Reinterpretation of the reflection seismic profile AR16-81 showing interpreted probable faults and top horizons corresponding to the distinct lithostratigraphic units.

Finally, the top horizons of Grés de Almargem and Barremian-Berriasian units in time were depth-converted using seismic velocities obtained from well data outside the study area for geological units similar to those found inside the region of interest.

3.3. Database enhancement

Finally, the database was refined, creating a single file using Microsoft Office Excel. Each point from different sources has associated coordinates of x, y and z, and horizon code (1 – top, 2 – bottom, 0.5 – relative top, 1.5 – minimum bottom). Thereafter a selection of the most representative data concerning the top of Grés de Almargem and Barremian-Berriasian units was performed (Table 3).

Table 3: Enhanced database summary.


This database, constituted by 6,035 points from the diverse data sources considered, is the potential input for the future three-dimensional geological modelling of the area (Figure 7).

Figure 7: Location of top and bottom points (grouped by horizons) obtained from the geological (C: cartography), hydrogeological (B: boreholes) and geophysical (S: seismic) data.

  1. Discussion and conclusion

This study has shown a methodology for integrating data from different sources into a single georeferenced database, which will allow the future construction of a three-dimensional geological model of the Lower Cretaceous units in the Lisbon region. Figure 8 presents a flowchart with the phases of the work.

Figure 8: Flowchart showing the distinct phases of the work.

In the eastern part of the study area, where a strongly demand for energy exists and predictably Lower Cretaceous units are deeper (and thus a higher temperature of groundwater is expected), there is still a scarce amount of data (Figure 7). Groundwater wells in this sector usually exploit shallower formations, namely the Miocene and Paleogene units, and there are no seismic profiles. However, seismic profiles previously reprocessed and reinterpreted by Carvalho et al. (2005; 2011) exist at Tagus estuary. Therefore, in order to complete the gaps, the next step would be a junction of those profiles with the data obtained in the present study.

Finally, it is important to note the dynamic feature of this methodology, which in the future may be improved with new data and adopted for other geosciences applications besides geothermal purposes (e.g. structural geology, hydrogeology, CO2 storage, etc.).


This study is part of the master’s final work of the first author, advised by Prof. Sofia Verónica Trindade Barbosa, who we thank for all the support received. This study is also a contribution to Projects SFRH/BPD/76404/2011 and UID/GEO/04035/2013 funded by FCT (Fundação para a Ciência e a Tecnologia) in Portugal. We express our sincere thanks to the Portuguese Environment Agency (APA) for allowing access to well reports, and the National Entity for the Fuel Market (ENMC) for its kindness in the provision of seismic and aeromagnetic information.


Carvalho, J.M., Cardoso, A.A.T. 1994. The Air Force Hospital geothermal project in Lisbon. In: Geothermics’94 in Europe, Orléans, France. Documents-BRGM. 230. 441-448.

Carvalho, J.Matias, H., Torres, L. Manupella, G., Pereira, R., Mendes-Victor, L. 2005. The structural and sedimentary evolution of the Arruda and Lower Tagus sub-basins, Portugal. Marine Petrology Geology. 22(3). 427–453. DOI: 10.1016/j.marpetgeo.2004.11.004

Carvalho, J., Rabeh, T., Bielik, M., Szlaiova, E., Torres, L., Silva, M., Carrilho, F., Matias, L. Miranda, J. M.. 2011. Geophysical study of the Ota–V.F. Xira-Lisboa-Sesimbra fault zone and of the Lower Tagus Cenozoic basin. Journal of Geophysics and Engineering. 8. 395-411. DOI: 10.1088/1742-2132/8/3/001

Correia, A., Ramalho, E., Rodrigues da Silva, A.M., Mendes-Victor, L.M., Duque, M.R., Aires-Barros, L., Santos, F.M., Aumento, F., 2002. Portugal. In: Hurter, S. and Haenel, R. (Eds.), Atlas of Geothermal Resources in Europe. GGA, Hannover. 47-49.

Dias, R., Araújo, A., Terrinha, P., Kullberg, J. C. (Eds.). 2013. Geologia de Portugal, Vol. II, Geologia Meso-cenozóica de Portugal (Geology of Portugal, Vol. II, Meso-cenozoic geology of Portugal), Escolar Editora, Lisboa. 195-347.

Marrero Diaz, R., Ramalho, E., Costa, A., Ribeiro, L.; Carvalho, J., Pinto, C., Rosa, D., Correia, A. 2015. Updated geothermal assessment of lower cretaceous aquifer in Lisbon Region, Portugal. In: Proceedings World Geothermal Congress, Melbourne, Australia. Extended abstract 16012.

Rasmussen, E.S., Lomholt, S., Andersen, C., Vejbaek, O.V.. 1998. Aspects of the structural evolution of the Lusitanian Basin in Portugal and the shelf and slope area offshore Portugal. Tectonophysics. 300. 199-225. DOI: 10.1016/S0040-1951(98)00241-8

Rey, J., 1993. Les unités lithostratigraphiques du Crétacé inférieur de la región de Lisbonne (The Lower Cretaceous lithostratigraphic units of the Lisbon region), Comunicações dos Serviços Geológicos de Portugal, 78(2), Lisboa. 103-124.

Walker, D., J., 1983. Final Report. Seismic Interpretation. Lusitanian basin, Portugal. Report No. 22391 for Petrogal, Lisboa.

This article has been published in European Geologist Journal 44 – Geology and a sustainable future.

Read here the full issue: