1. Introduction

Historical population grids are scarce or rather nonexistent. This work represents a first –but maybe not the last– effort in this direction1. Using historical cadastral data and homogeneous population data at municipal level, constructed in different projects, we generate, for the whole of Spain, population grids with 100m x 100m and 1km x 1km resolutions and all census years from 1900 to 2021.

These grids are top-down (Goerlich and Cantarino 2014). The methods used are similar to those employed in the generation of population grids for the entire globe in the last decades by combining information from satellite images with demographic information from censuses (Schiavina et al 2023).

Given the richness of cadastral information (Zoraghein, Leyk, Ruther and Buttenfield 2016), and the possibility of going much further back in time than satellite information, we can generate much older population grids. Although far from perfect, these grids provide a much better approximation to the distribution of the population in those years than the simple consideration of the municipal population or the count in the settlements derived from the gazetteers associated with the censuses.

The possibility of taking our estimates up to 2021, where we have a bottom-up population grid from the Spanish National Statistical Institute (INE), derived from the 2021 census, allows us to validate our methods, albeit for the most recent dates.

We cover the total boundary of Spain, including Canary Islands, as shown in map 1.

Map 1: Study area
Spain and inhabited cells, 1km x 1km, of the estimated population grid for 2021 (HIPGDAC-ES).

Source: HIPGDAC-ES, IGN and own elaboration.

The structure of the work is as follows. The data sets used, as well as their advantages and disadvantages, are described in detail below. Section 3 describes the methods, and indicates the various products generated. The next section briefly reviews the results obtained for the 1km x 1km grids and the entire period considered. Section 5 performs a tentative validation exercise with the limited information available. Finally, the paper concludes with a brief reflection.

2. Original information

Although the main sources of information for the generation of the top-down population grids are a historical compilation of cadastre data (Uhl et al 2023) and the homogenised census population data at municipal level –Local Administrative Units (LAU)– (Goerlich et al 2015), the work uses other complementary sources for quality improvement. This section describes in detail the different databases used.

2.1 Historical settlement data compilation for Spain (1900-2020): HISDAC-ES

Uhl et al (2023) download and process the cadastral information of more than 12 million buildings and their features. After a certain harmonisation of the different existing cadastres in Spain –five in total–, they transform the vector layers of the building contours into point vector information, and the different features are transformed into raster layers with 100m x 100m resolution. The database is labeled as HISDAC-ES2. Validation exercises carried out by Uhl et al (2023) indicate that the database is of acceptable quality, especially in recent times, probably higher than remote sensing products, although at the beginning of the century the quality of the information is considerably reduced due to the problem of dating the buildings in the cadastre, the so-called survivorship bias.

The volume of information generated is enormous –743 raster layers– and many of the variables provided have been classified by five-year intervals since 1900. The raster information is offered in different Coordinate Reference Systems (CRS), including ETRS89-LAEA3 –EPSG:3035–, which is the reference system for European grids (INSPIRE 2014), for all Spanish territory.

This information constitutes the main support on which the population will be allocated. Of all the variables offered, the most appropriate for redistributing the resident population by administrative unit over the territory –the grid cells with a resolution of 100m x 100m– is total building indoor area (RES_BIA), which is available for each decade from 1900 to 2020 –approximate in line with the decennial census–4. This variable includes the two most relevant factors in the processes of spatial disaggregation of the population (Goerlich 2016): the residential function and an indicator of the height of the buildings.

However, cadastral information is not uniform in space and time, and this affects the quality of the population redistribution.

From the spatial point of view, the existence of 5 cadastres for the whole territory, one for the majority of Spain, and 4 for each of the provinces under a different tax system from the rest of the Spanish territory –the 3 provinces of the Basque Country and Navarre– affects the uniformity of the available information.

For example, RES_BIA is not available in any year for the 3 Basque provinces –Araba/Álava, Gipuzkoa and Bizkaia–. Simply this variable is not in the data model of the cadastres of the Basque Country. For Araba/Álava we only have information on residential building footprint area (RES_BUFA), so that we lose information about the height of the buildings. And for Gipuzcoa and Bizkaia we do not even have information on the use of the building, so that only the building footprint area (BUFA) is available. As a result, the disaggregation process should incorporate information from these two variables, RES_BUFA, and BUFA, in some cases.

As mentioned at the beginning of the section, if we go back in time, things get worse. As recognized by Uhl et al (2023), the HISDAC-ES database has some deficiencies in its early years. This situation is not uniform in space and affects some places more than others. The main problem is the so-called survivorship bias. The HISDAC-ES database infers the age of the building from the date given in the cadastral information, but this date may not correspond to the initial date of construction of the building. The reason is that cadastre changes the date of construction of a building if there has been a restoration or modification of the building of a certain entity, and does not keep record of these changes –at least in the available public information–. As a result, it remains unknown whether a construction date represents the truly first establishment of a building at a given location or whether there was a similar built-up structure existing prior to that. The problem is noted by Uhl et al (2023), section 4.6, which indicate how it is especially evident for some municipalities at the beginning of the 20th century, and urge users to take precautions when analyzing the data in the long term.

Additionally, minor issues warrant consideration. To illustrate, we may cite the case of buildings that existed in the past but have since been demolished. Such buildings are no longer included in the cadastre and, therefore, are not reflected in HISDAC-ES. Furthermore, the function of a given building within the cadastre may be its current function, although it is possible that this function may have differed in the past. Furthermore, it is unclear whether all the characteristics of a building are based on the same date of measurement or if they were recorded on the date of construction, which may differ from the date of measurement. As previously stated, HISDAC-ES is the product of the integration of five distinct cadastres, each with its own data model. Consequently, inconsistencies may exist between the data sets. Indeed, it has already been stated that certain crucial variables are not accessible in all instances. Furthermore, the treatment of buildings as point elements, rather than polygons, and their management as raster layers, rather than with the original vector information, introduces certain distortions. Given the magnitude of the database, these distortions are necessary to reduce the computational burden.

In any case, all available evidence suggests that these are relatively minor issues in comparison to the survivorship bias. However, it is essential to consider them when evaluating the results and to recognise that they should diminish over time, given that they currently have a lesser impact on the available information than they did at the beginning of the 20th century.

To get an idea of the importance of the survivorship bias in extreme cases we can examine which municipalities lack cadastral information in the different census years of the 20th century. If we intersect the municipal boundaries from the National Geographical Institute (IGN) with the layers of RES_BIA, RES_BUFA and BUFA of HISDAC-ES, we will observe that none of the municipalities have zero intersection only from 1970. Even in 1960, there is a municipality that does not have information about buildings in HISDAC-ES. It is a small municipality, Beizama (20020) in the province of Gipuzkoa, with only 460 inhabitants. Of course, this problem gets worse as we go back in time. In 1950, this lack of information affected 4 municipalities, 15 in 1940, 29 in 1930, 57 in 1920, 111 in 1910 and 171 in 1900. In all cases, except for a minor exceptions in the first decades of the 20th century, these are very small municipalities.

It should be noted that this is a general problem that affects all of the municipalities, although it manifests itself in an extreme way in some more than others, and is also present in large cities. Natural aggregation to cells, especially 1km x 1km, dilutes this problem and smooths the survivorship bias, although quantification of the problem is extremely difficult.

A problem related to the survivorship bias, not mentioned in Uhl et al (2023), has to do with the dating of many rural isolated farmhouses whose cadastre date does not really correspond to the date of construction. Many of these buildings were inhabited at the beginning of the 20th century and gradually ceased to be so throughout the second half of that century. The indications of this incorrect dating are evident from the fieldwork and direct consultation to the cadastre data of certain buildings. The problem in some places is so obvious that a direct and personal consultation on this matter was made to the rural section of the cadastre. The answer is particularly revealing:

“…it must be taken into account that the collection of information on the dates of construction of buildings on rural land, was carried out for most of the national territory in a massive way in the 1990s.

This work was contracted to external companies which made a construction sheet that reflected among other features, its use and an estimated date of seniority.

The nature of the fieldwork carried out, as well as the enormous task of covering the entire national territory, sometimes prevented the exact date of construction from being known, in such a way that it is not possible to guarantee the usefulness of this data to be used in publications of a scientific nature that may require more rigour.” (e-mail of 28/04/2023)5

There is no apparent solution to this type of problem, but it is necessary to be aware of it in order to be able to evaluate the results obtained.

What is obvious is that the information of HISDAC-ES must be complemented with another, especially in the first decades of the 20th century, if we start from municipal information to generate historical population grids.

2.2 Historical municipal population data: 1900 - 2021

The main source of population information is the homogeneous database of Goerlich, Ruiz, Chorén and Albert (2015). These authors generate estimates of municipal population with the structure of municipalities from the 2011 census, so that they take into account the municipal alterations –mergers and segregations, total and partial– occurred between 1900 and 2011. We therefore have homogeneous municipal populations for the 8,116 municipalities that appear in the 2011 census. In addition, to reach the most recent date, the database is expanded with the populations of the 8,131 municipalities of the 2021 census.

The contours of the municipalities come from the National Geographical Institute (IGN) with reference dates 01/01/2012, for the period 1900 - 2011 –8,116 municipalities–, and 01/01/2019, for 2021 –8,131 municipalities–. Since the IGN municipal enclosures and boundary limit lines database is continuously updated and the IGN does not maintain historical data, these are collected and organized by Goerlich and Perez (2021).

2.3 Settlement coordinates from 1887 census: ESPAREL

As has become clear in the description of HISDAC-ES at the beginning of this section, we must incorporate external information that mitigates, as far as possible, the survivorship bias.

An independent geocoding project of the historical population entities contained in the gazetteers of the late eighteenth and nineteenth centuries is ESPAREL (Beltrán Tapia, Díez Minguela, Fernández Modrego, Gómez Tello, Martinez-Galarraga and Tirado Fabregat 2022)6. The main objective of ESPAREL has been to generate a spatial data infrastructure (SDI) that allows linking the spatial planning of the Old Regime with that of the liberal State at the end of the 19th century, and at the same time with the current one. For this purpose, the existing population entities in the 1787 census –census of Floridablanca–, and those of the 1887 Spanish Gazetteer –corresponding to the census of the same year– have been linked with the Basic General Gazetteer of Spain (NGBE), which is geocoded, as well as other external geocoded information currently available from regional cartographic institutes or map viewers accessible through external queries from APIs.

From our point of view, what matters is that the gazetteer of 1887 is relatively close to 1900, which is our first year for the generation of historical population grids from HISDAC-ES, and for which we have the point coordinates of 58,837 population entities, covering almost all of the current municipalities. We incorporate this information for those municipalities in which HISDAC-ES lacks it. This solves all the cases mentioned above except those of 4 newly created municipalities, for which the homogeneous populations of Goerlich, Ruiz, Chorén and Albert (2015) indicate the existence of population, but not cadastre –HISDAC-ES–, neither ESPAREL gives geocoded information that allows to bound in the space the population of these municipalities7.

It should be remembered that the geocoded information of ESPAREL does not refer to buildings, but to the population entity itself. However, the number of buildings and their plants are available in the database and, by construction, they are residential buildings. From this information a variable was constructed which adds the buildings once multiplied by their plants, which in a certain way represents the closest to the variable RES_BIA of HISDAC-ES, except that the coordinates do not correspond to individual buildings, but to the population entity. This vector point information was transformed into a raster layer with the same structure as the HISDAC-ES layers. Note that each entity will belong to a single cell of 100m x 100m, regardless of its size, but since these are entities of low demographic weight this is not expected to generate many distortions, especially on cells added to a resolution of 1km x 1km.

As a last resort, for the municipalities for which we do not find information either in ESPAREL, we use the coordinate of the capital of the municipality, which is available in the Gazetteer of Municipalities and Population Entities (NGMEP) of the IGN. Like the coordinates of ESPAREL this vector point information was transformed into a raster layer with the same structure as those of HISDAC-ES, and the population of the municipality assigned to the corresponding cell. The small size of the population in these few case makes distortions minimal.

2.4 Other information used

In addition to the above information, used in the process of downscaling the municipal population to the grid, the Global Human Settlement Layer (GHSL) grid population data (Schiavina, Freire, Carioli and MacManus 2023), version R2023A (Schiavina et al 2023), with resolution of 1km x 1km, were used for the years 1980, 1990, 2000, 2010 and 2020 to compare their results with those obtained by our methods. The GHSL is a worldwide product that had to be masked by Spain’s administrative boundaries for comparison with our results.

Finally, we also used, for comparison purposes, the INE census grid derived from the 2021 census, GEOSTAT2021, a grid generated by bottom-up methods.

3. Downscaling historical municipal population data: Methods

The methods are relatively simple, and are not very different from those used in the production of the GHSL population grids (Freire, MacManus, Pesaresi, Doxsey-Whitfield and Mills 2016).

The approach is based on the well known vector/raster based dasymetric mapping (Wright 1936, Mennis 2003) relying mainly on the variable RES_BIA from HISDAC-ES to restrict and refine the distribution of people. Even in 1900, this covers 93% of municipalities, that increases to 97% in 2021. When, for a given municipality, this information is insufficient or non-existent other less appropiate variables are used following a clear predefined order. Figure 1 shows the workflow for population disaggregation. The process is carried out at the municipal level, so for each municipality the best support for the population is sought in the available information, in order to aggregate the results later.

Figure 1: Workflow of population disaggregation model.

Source: Own elaboration.

For each period, population grids were produced following a clear and concise volume-preserving dasymetric mapping approach at LAU level. So, for a given period we iterate over municipalities looking for population support in the sequential order shown in Figure 1. Given a vector census layer of municipal boundaries and a HISDAC-ES raster layer, for a populated municipal polygon (source zone) we proceed as follows:

  1. If a particular municipality has information on the total building indoor area (RES_BIA), then we downscale population proportionally to RES_BIA.

  2. If there is no information on RES_BIA, but the municipality has residential building footprint area (RES_BUFA), then we downscale population proportionally to RES_BUFA.

  3. If there is no information on the previous variables, but the municipality has building footprint area (BUFA), then we downscale population proportionally to BUFA.

  4. If there is no information on HISDAC-ES for a given municipality, then we use the ESPAREL coordinates of population entities to redistribute the municipal population.

  5. As a last resort, if no other information is found, in neither HISDAC-ES or ESPAREL, we assign municipal population to the coordinates of the municipal capital.

This sequential process ensures that, for each municipality, we use the best information available at each point in time. Steps 4 and 5 are necessary because, for some municipalities, at the beginning of the 20th century we did not find support for the population in HISDAC-ES. It should be obvious that the process is carried out for each year and municipality sequentially.

Table 1, below, shows the percentage of municipalities and population by type of support on which to locate the population. Overall, the results are reasonably good. In 1900, 93% of the municipalities and 95% of the population find support in the most reliable variable, RES_BIA, and the percentage of municipalities or population for which we must use auxiliary information, as we do not find support for it in HISDAC-ES, is very low, reaching only 1% of the population in 1900, when we have to use the settlement coordinates from 1887. The results improve significantly over time, and from the middle of the 20th century onwards it is no longer necessary to use information external to HISDAC_ES. The results due to the variable RES_BUFA come from the 51 municipalities of Araba/Álava, whose land registry does not provide information on the building indoor area.


Table 1. Municipalities and population shares (%) by support and census year.
Municipalities (%)
Population (%)
Year RES_BIA RES_BUFA BUFA 1887 Capital RES_BIA RES_BUFA BUFA 1887 Capital
1900 92.6 0.63 4.66 2.06 0.05 94.6 0.52 3.77 1.0847 0.0017
1910 93.6 0.63 4.36 1.32 0.05 95.2 0.49 3.77 0.5219 0.0019
1920 95.2 0.63 3.46 0.65 0.05 95.8 0.46 3.54 0.1871 0.0026
1930 96.1 0.63 2.88 0.31 0.05 96.1 0.44 3.39 0.0779 0.0034
1940 96.5 0.63 2.72 0.14 0.05 96.3 0.43 3.22 0.0158 0.0039
1950 96.8 0.63 2.48 0.04 0.01 96.3 0.41 3.29 0.0034 0.0040
1960 96.9 0.63 2.48 0.01 95.6 0.43 3.98 0.0015
1970 96.9 0.63 2.46 94.5 0.59 4.90
1981 96.9 0.63 2.46 94.3 0.68 5.00
1991 96.9 0.63 2.46 94.6 0.70 4.71
2001 96.9 0.63 2.46 94.9 0.70 4.40
2011 96.9 0.63 2.46 95.3 0.69 3.98
2021 96.9 0.63 2.46 95.3 0.70 3.96
Source: HISDAC-ES, Census 1900 - 2021 (INE) and own elaboration.


In fact, if we had to identify areas with insufficient or inadequate representation, we would find them mostly in the Basque Country. Figure 2 shows the estimated inhabited cells, with a resolution of 1km x 1km, for the provinces of Bizkaia and Gipuzkoa in the census years 1900, 1940, 1981 and 2021, and we observe two different situations where the spatial information supporting the population distribution generates an inadequate and different contribution. These two cases represent the two most extreme situations we found. On the one hand, the information for Bizkaia comes entirely from the footprint of the buildings (BUFA), without classification by use, which generates an excessive dispersion of the population visible in all the years, especially in 1900. On the other hand, the information for Gipuzkoa is much more deficient, and in 1900 it comes largely from settlements coordinates of the of 1887 census, which generates an excessive concentration of the population in a few cells at the beginning of the 20th century. This phenomenon is also observed in 1940 and gradually disappears as the 20th century progresses. It is important to remember that the cadastres of these two provinces are different, in the data model and in the informational content.

Figure 2: Inhabited cells, 1km x 1km, of the estimated population grid for 1900, 1940, 1981 and 2021 and the provinces of Bizkaia and Gipuzkoa (HIPGDAC-ES)

Source: HISDAC-ES, Census 1900 - 2021 (INE), ESPAREL and own elaboration.

These situations are not generally observed in other regions. Figure 3 shows the same information as Figure 2 for Madrid and allows us to observe a more realistic evolution in the populated cells. In this case, for almost all the municipalities we have information on the total building indoor area (RES_BIA) as early as 1900. Most regions show a pattern like that of Figure 3, so the results seem similar in quality to HISDAC-ES, and although the overall results seem very plausible, there are areas where the final data produced may be more deficient. The population distribution algorithm stores intermediate information that allows tracking quality at the municipality level.

Figure 3: Inhabited cells, 1km x 1km, of the estimated population grid for 1900, 1940, 1981 and 2021 and the province of Madrid (HIPGDAC-ES)

Source: HISDAC-ES, Census 1900 - 2021 (INE), ESPAREL and own elaboration.

The vector/raster dasymetric mapping approach is implemented on the original resolution of the HISDAC-ES raster layers, 100m x 100m, and the population per cell rounded to integer values at this resolution using LAU population as totals –except for the 2011 census, ¡whose population values published by INE are real, 👎!–. Hence, population grids are integer valued at this resolution, and aggregated afterwards to the standard resolution of 1km x 1km, in raster and vector formats. The integer adjustment algorithm, performed at the 100m x 100m cell level, is the method of largest remainder (Cox 1987; Balinski and Rachev 1993, 1997). The integer adjustment is a novelty, that has traditionally benn overlooked in the literature on dasymetric population disaggregation.

Table 2, below, shows the population of each census and the number of inhabited cells estimated with the above algorithm. According to our estimates, while the population has multiplied by 2.5 in the period 1900 - 2021, the inhabited cells have only doubled.


Table 2. Census population and estimated inhabited grid cells.
Year Population Cells
1900 18,830,649 71,804
1910 20,360,306 74,200
1920 22,012,663 80,564
1930 24,026,571 86,375
1940 26,386,854 93,203
1950 28,172,268 101,162
1960 30,776,935 108,718
1970 34,041,482 116,504
1981 37,682,355 127,492
1991 38,872,268 133,701
2001 40,847,371 138,786
2011 46,815,916 148,888
2021 47,400,798 143,457
Source: HISDAC-ES, Census 1900 - 2021 (INE)
and own elaboration.


A close inspection of table 2 reveals some striking figures that deserve a few comments. The number of inhabited cells continuously increases up until the last two years, 2011 and 2021. In this case the last census, 2021, shows a significant drop in the number of inhabited cells compared to the previous census, 2011. One might think that this trend is due to the huge population growth between 2001 and 2011, as opposed to the low population growth between 2011 and 2021. However, this is not the case. It is easy to show that the result is, in a way, a statistical artifact derived from the fact that the INE published population data in real figures for the 2011 census, which as INE (2023) itself acknowledges was not a good idea, and introduced additional complications. Our estimates maintain the INE population structure, and therefore the layers generated provide population figures in integers at the 100m x 100m cell level for all years except 2011, where the results of the disaggregation process are not rounded. The result is that the population in real figures overestimates the number of inhabited cells. If for 2011 we round off the population figures to integers at the 1km x 1km cell level8, then the population obtained is 46,815,735 persons, slightly different from the total census population –table 2–, and the number of inhabited cells would be 143,974, much more in line with the trend observed in table 2. Even in this case the number of inhabited cells would be slightly lower in 2021 than in 2011, but given the magnitude of the differences we cannot rule out that this would not be the case if the rounding were done at the level of 100m x 100m cells.

Another curious fact, which is not possible to observe only with the information in table 2, is that the trend observed for the years 2011 and 2021 in our estimates is radically different from that observed when we compare the only two population grids published by INE, derived by bottom-up procedures, and which correspond to the 2011 and 2021 censuses, where the inhabited cells in 2021 practically double those that existed in 2011 (Goerlich 2024). As indicated in Goerlich (2024), section 5, this is due to a design effect in the generation of the 2011 grid, which makes it unreliable for determining the actual number of inhabited cells in Spain in that year.

As a result of this process, we have 13 raster population layers with resolution of 100m x 100m –GeoTiff format–, 13 raster population layers with resolution of 1km x 1km –GeoTiff format–, and a vector file with all populated cells in at least one census year –GeoPackage format (gpkg)–, with resolution of 1km x 1km, as well as the cell indicator according to the INSPIRE (2014) directive. This information is public and freely available from zenodo and GitHub.

The whole process was performed using free software based on the statistical computing system R (R Core Team 2023), version 4.4.1, using tidyverse libraries (Wickham et al 2019) for data wrangling, version 2.0.0, sf library (Pebesma 2018) for handling vector information, version 1.0-17, and terra (Hijmans 2023), version 1.7-78, and stars (Pebesma and Bivand 2023), version 0.6-6, libraries for handling raster information.

4. Historical population grid data at 1km x 1km resolution: 1900 - 2021

This section illustrates some of the results obtained from the 1km x 1km resolution population grids. In a way, it has a dual purpose. On the one hand, it is a first descriptive analysis of the historical evolution of the population distribution in this format. On the other hand, it serves as a first tentative validation exercise, in the sense of examining whether the results obtained are in agreement with the main known historical trends in the evolution of the distribution of the Spanish population throughout the 20th century and the beginning of the 21st century. These trends are well known (Goerlich and Mollá 2021).

First, table 3 shows the distribution of cell sizes in relative terms over the period 1900 - 20219. Large trends are observed at the extremes of the distribution. Thus, cells with a small population, up to 10 inhabitants, go from representing 12% in 1900 to more than tripling in 2021, representing 42%. At the other extreme, cells with more than 1,000 inhabitants also increase their relative weight, although only those with more than 10,000 inhabitants show a marked trend. In this case, their relative importance is low, not even reaching 1% in 2021, but their weight has increased fivefold in these 120 years. The remaining cell sizes, above 25 and up to 1,000 inhabitants, show a decreasing relative importance, especially in the interval between 100 and 200 inhabitants.


Table 3. Distribution of cells (%) by population size.
Interval 1900 1910 1920 1930 1940 1950 1960 1970 1981 1991 2001 2011 2021
Up to 10 12.04 11.96 13.06 14.52 16.57 19.57 23.23 29.16 34.81 37.80 40.04 42.62 42.23
(10, 25] 14.97 14.80 15.45 15.95 16.38 16.90 17.07 17.00 16.86 16.71 16.30 15.85 15.90
(25, 50] 15.57 15.27 14.93 14.93 14.75 14.52 14.07 13.76 12.96 12.39 12.01 11.15 11.14
(50, 100] 17.17 16.94 16.90 16.41 15.84 15.06 14.03 12.90 11.56 10.77 9.91 8.96 8.91
(100, 200] 16.05 16.11 15.78 15.15 14.35 13.28 12.27 10.53 9.04 8.22 7.53 6.76 6.65
(200, 300] 7.51 7.71 7.43 6.95 6.56 6.05 5.41 4.44 3.82 3.44 3.21 3.05 3.07
(300, 500] 6.69 6.86 6.45 6.29 5.89 5.40 4.94 4.13 3.54 3.25 3.18 3.08 3.14
(500, 1000] 5.45 5.51 5.24 5.01 4.90 4.58 4.22 3.57 3.12 3.06 3.06 3.13 3.25
(1000, 5000] 4.08 4.35 4.23 4.21 4.08 3.95 3.97 3.55 3.25 3.30 3.59 4.05 4.27
(5000, 10000] 0.34 0.35 0.37 0.40 0.43 0.46 0.50 0.51 0.52 0.53 0.60 0.75 0.80
(10000, 20000] 0.08 0.10 0.10 0.12 0.14 0.14 0.18 0.28 0.31 0.32 0.36 0.41 0.44
More than 20000 0.05 0.05 0.06 0.07 0.09 0.09 0.12 0.16 0.22 0.21 0.19 0.18 0.19
Total 100.00 100.00 100.00 100.00 100.00 100.00 100.00 100.00 100.00 100.00 100.00 100.00 100.00
Source: HISDAC-ES, Census 1900 - 2021 (INE) and own elaboration.


Second, table 4 shows the percentage of inhabited cells out of the total at the Autonomous Community level, i.e. the proportion of inhabited area with a resolution of 1km x 1km. Naturally, given the increase in population, all regions show increasing percentages of human occupation, but in some regions the trend is particularly striking. Thus, the Community of Madrid multiplies its occupied area by 6.4, Extremadura by 4.8, Andalusia by 4.2 and the Region of Murcia by 2.5. The results also show, clearly, the regions with the most dispersed population –in the sense of showing greater human occupation of their territory–, Galicia, Asturias and the Basque Country. In 2021, with the exception of Ceuta and Melilla, only Galicia exceeds 60% of populated territory, although the Balearic Islands, the Basque Country and the Principality of Asturias exceed 50%, whereas the Region of Murcia is right in this percentage.


Table 4. Share (%) of inhabited cells by Autonomous Community.
Code Region 1900 1910 1920 1930 1940 1950 1960 1970 1981 1991 2001 2011 2021
01 Andalucía 7.2 7.9 9.7 11.8 14.5 17.8 20.5 22.6 25.2 26.7 28.3 30.9 30.0
02 Aragón 5.8 6.0 6.3 6.6 7.1 7.7 8.2 8.8 10.0 10.6 10.8 12.0 11.1
03 Principado de Asturias 47.6 48.2 48.8 49.2 49.6 50.5 51.6 51.9 52.1 52.2 52.2 53.9 51.8
04 Illes Balears 31.6 32.2 36.0 37.5 39.5 41.1 43.1 46.6 50.0 51.8 53.2 54.9 54.6
05 Canarias 21.8 22.6 24.3 25.9 27.7 30.3 33.6 36.4 40.1 42.7 45.8 47.8 47.7
06 Cantabria 25.8 26.4 34.0 38.0 39.0 39.8 40.3 41.1 41.6 42.4 42.6 43.7 42.8
07 Castilla y León 9.0 9.3 10.3 10.8 11.3 11.8 12.3 12.8 13.9 14.6 15.1 16.2 15.3
08 Castilla-La Mancha 4.3 4.5 5.1 5.7 6.4 7.2 8.0 9.0 11.1 12.4 13.7 15.9 14.7
09 Cataluña 30.8 31.6 32.5 33.4 34.4 36.2 37.9 40.5 43.5 45.1 46.1 48.3 47.2
10 Comunidad Valenciana 19.7 20.8 23.0 25.6 27.9 30.4 33.6 38.3 43.4 45.6 46.7 50.1 47.5
11 Extremadura 3.7 3.9 4.7 5.7 7.8 9.6 11.3 12.5 14.3 15.4 16.6 19.5 17.8
12 Galicia 54.6 55.6 57.4 58.7 59.7 60.5 61.2 62.0 63.1 64.1 64.6 65.7 64.8
13 Comunidad de Madrid 6.4 6.8 7.4 8.5 10.2 12.6 17.0 23.6 33.0 36.1 38.2 41.3 41.2
14 Región de Murcia 20.2 21.7 26.1 29.5 32.8 38.0 40.9 43.1 45.6 47.2 48.4 50.3 50.0
15 Comunidad Foral de Navarra 11.0 11.1 11.3 11.5 11.8 12.2 12.8 13.6 15.0 15.9 16.7 17.7 17.9
16 País Vasco 35.9 35.4 36.7 38.0 39.4 41.3 43.3 45.9 49.7 51.4 53.0 54.9 54.5
17 La Rioja 8.7 8.9 9.3 9.6 10.2 11.1 12.0 13.8 15.7 16.7 17.2 19.3 17.5
18 Ceuta 15.4 15.4 23.1 35.9 41.0 51.3 59.0 59.0 64.1 64.1 64.1 64.1 64.1
19 Melilla 11.1 19.4 30.6 36.1 36.1 38.9 41.7 41.7 41.7 41.7 47.2 47.2 47.2
Total 14.0 14.5 15.8 16.9 18.2 19.8 21.3 22.8 24.9 26.1 27.1 29.1 28.1
Source: HISDAC-ES, Census 1900 - 2021 (INE) and own elaboration.


Two well-known trends in the evolution of population distribution throughout the 20th century is its tendency to move towards the valley and towards the coast (Goerlich and Mollá 2021)10. Using data from Goerlich (2023) it is possible to quantify these trends from the generated population grids.

Table 5 shows the evolution of population distribution by altimetric cuts for the period 1900 - 2021, where altitude is measured, as is population, at cell level (Goerlich 2022). For comparability, the same altimetric cuts are used as in Goerlich and Mollá (2021), and broadly very similar trends are observed. If the resident population below the 200 meters range was about a third in 1900, in 2021 it reached 53%. The rest of the intervals decreased their relative importance, with the strip above 1,000 meters of altitude showing the biggest decrease.


Table 5. Population share (%) by altitude (m).
Altitude 1900 1910 1920 1930 1940 1950 1960 1970 1981 1991 2001 2011 2021
Up to 200m 33.5 34.1 35.2 36.3 37.8 39.1 41.7 46.5 49.8 50.9 51.6 52.4 53.0
(200, 600] 30.6 30.5 30.1 29.4 28.5 27.9 26.9 24.7 23.2 22.7 22.5 22.3 22.0
(600, 1000] 30.2 29.9 29.7 29.6 29.4 29.0 27.9 26.4 25.2 24.7 24.4 23.9 23.7
More than 1000m 5.7 5.4 5.1 4.7 4.3 4.0 3.5 2.4 1.8 1.6 1.5 1.4 1.3
Total 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0
Source: HISDAC-ES, Census 1900 - 2021 (INE) and own elaboration.


Finally, Table 6 shows the evolution in the distribution of the population according to its distance to the coast in different intervals11. Distance, like population, is measured at cell level (Goerlich 2023).

Clearly, Spain was already a coastal country in 1900 (Rappaport and Sachs 2003), since a quarter of its population was located on a 10km strip off the coastline, but in 2021 this percentage reached 40%. The rest of the intervals, between 10 and 200km, have seen their relative population decrease, and the slight rise in population observed at more than 200km of the coastline, 3 percentage points, is undoubtedly due to the increase of the metropolitan area of Madrid (Goerlich and Mollá 2021).


Table 6. Population share (%) by distance to the coast (km).
Distance 1900 1910 1920 1930 1940 1950 1960 1970 1981 1991 2001 2011 2021
Up to 10km 24.5 25.1 25.7 26.8 27.9 28.8 30.6 34.7 37.2 37.8 38.2 39.0 39.5
(10, 50] 22.8 22.3 21.7 20.8 20.0 19.4 19.2 18.6 18.4 18.3 18.4 18.5 18.6
(50, 100] 15.8 15.7 15.6 15.2 15.0 14.8 13.9 12.5 11.7 11.6 11.3 10.9 10.7
(100, 200] 19.6 19.4 19.3 19.1 18.8 18.4 17.1 14.5 13.1 12.7 12.1 11.3 10.7
More than 200km 17.4 17.5 17.7 18.1 18.2 18.6 19.2 19.6 19.6 19.7 20.0 20.3 20.5
Total 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0
Source: HISDAC-ES, Census 1900 - 2021 (INE) and own elaboration.


On the whole, the results seem reasonable and show the major trends already known about the evolution in population distribution throughout the 20th and early 21st century, measured at the municipal level (Goerlich and Mollá 2021), although we are far from a real validation exercise, but at least this shows that the estimated grids makes sense.

5. Comparison with other population grids

5.1 Global Human Settlement Layer Population grids

The Global Human Settlement Layer (GHSL) publishes, in its version R2023A, population grids between 1975 and 2030 with a five-year periodicity12, resolutions of 100m x 100m and 1km x 1km and the Mollweide projection13. We can therefore compare our results for the years 1981, 1991, 2001, 2011 and 2021 with the corresponding GHSL results for 1980, 1990, 2000, 2010 and 2020. After all, our disaggregation methods are similar to those used by the GHSL (Freire, MacManus, Pesaresi, Doxsey-Whitfield and Mills 2016), although our support where to place the population differs substantially. In our case, the background information is primarily sourced from the cadastre. In contrast, the GHSL utilises raster layers of built-up areas derived from satellite imagery, categorised according to their functional classification (residential versus non-residential) and incorporating an estimate of the built-up volume through building height.

The comparison is made only from the 1km x 1km resolution. Once the tessellations corresponding to Spain have been downloaded and extracted, we join them together and cut them by the IGN administrative contours identical to those used for the generation of the population grid, once all the information has been transformed to the same coordinate reference system (CRS). The GHSL provides population figures per cell in real terms –and we have already observed that this is not a trivial matter when determining the number of inhabited cells–. To be consistent with our estimates we round cell-level population figures to integers, with the exception of 2010, as the 2011 census population figures are real, so we must retain this feature for comparison purposes.

Table 7 compares the population and inhabited cells from our estimates –table 2– with what we get from the GHSL from the previous process. The populations are reasonably similar, but the number of inhabited cells is not. The GHSL shows a much higher number of inhabited cells than our estimates, in the order of twice as many inhabited cells. All indications are that GHSL estimates disperse the population excessively. For 2011 we observe, in the GHSL estimates, the same effect as in HIPGDAC-ES related to the actual population. In fact, if we round up the GHSL population in 2011 at cell level the number of inhabited cells decreases to 265,081, slightly above 10%.


Table 7. Population and estimated inhabited grid cells:
HIPGDAC-ES versus GHSL.
HIPGDAC-ES
GHSL
Year Population Cells Year Population Cells
1981 37,682,355 127,492 1980 37,539,291 261,795
1991 38,872,268 133,701 1990 38,941,450 263,393
2001 40,847,371 138,786 2000 40,796,668 261,898
2011 46,815,916 148,888 2010 46,634,491 298,110
2021 47,400,798 143,457 2020 47,424,225 262,423
Source: HIPGDAC-ES, Census 1900 - 2021 (INE), GHSL2023A
and own elaboration.


Table 8 displays the distribution according to cell sizes in relative terms for the years that are common between HIPGDAC-ES and GHSL. The differences here are quite remarkable. GHSL estimates do not show a definite trend, but a high stability, unlike the estimates of HIPGDAC-ES. The proportion of cells of very small size, up to 10 inhabitants, is significantly higher in GHSL than in HIPGDAC-ES, and exceeds 50% in all years. The opposite is true for larger cells, where GHSL always offers less relative importance. This minor relative importance is actually perceptible in all cell sizes above 25 inhabitants.

To compare the similarity between percentage structures we can calculate the following discrepancy statistic:

\[\delta = \frac{1}{2}{\Sigma _j}\left| {{S_{{\rm{1}}j}} - {S_{{\rm{2}}j}}} \right| \tag{1}\]

where \(S_1\) and \(S_2\) represent the percentage structures under comparison.

This statistic, \(\delta\), ranges between 0, if both percentage structures are identical, and 1, in the case of maximum discrepancy, when percentage structures do not overlap.

The value of \(\delta\), expressed in percentage terms, \(100 \times \delta\), ranges from 17.2 in 1981 to 9.8 in 2021, with a monotonous downward trend. This convergence is clearly due to the tendency observed in the evolution of the percentages of HIPGDAC-ES, showing a clear growth in time of the smaller cells, which generates an approximation of the percentage structure of HIPGDAC-ES to the percentage structure of GHSL.


Table 8. Distribution of cells (%) by population size: HIPGDAC-ES versus GHSL.
HIPGDAC-ES
GHSL
Interval 1981 1991 2001 2011 2021 1980 1990 2000 2010 2020
Up to 10 34.81 37.80 40.04 42.62 42.23 51.64 51.04 51.85 54.51 51.48
(10, 25] 16.86 16.71 16.30 15.85 15.90 17.19 17.37 16.82 15.50 16.47
(25, 50] 12.96 12.39 12.01 11.15 11.14 10.76 10.92 10.52 9.64 10.19
(50, 100] 11.56 10.77 9.91 8.96 8.91 8.17 8.20 7.87 7.35 7.73
(100, 200] 9.04 8.22 7.53 6.76 6.65 5.14 5.17 5.06 4.92 5.16
(200, 300] 3.82 3.44 3.21 3.05 3.07 1.88 1.90 1.92 1.94 2.09
(300, 500] 3.54 3.25 3.18 3.08 3.14 1.51 1.56 1.66 1.74 1.93
(500, 1000] 3.12 3.06 3.06 3.13 3.25 1.42 1.45 1.65 1.67 1.87
(1000, 5000] 3.25 3.30 3.59 4.05 4.27 1.72 1.79 2.03 2.11 2.36
(5000, 10000] 0.52 0.53 0.60 0.75 0.80 0.32 0.33 0.36 0.36 0.40
(10000, 20000] 0.31 0.32 0.36 0.41 0.44 0.19 0.19 0.21 0.20 0.23
More than 20000 0.22 0.21 0.19 0.18 0.19 0.06 0.07 0.06 0.06 0.09
Total 100.00 100.00 100.00 100.00 100.00 100.00 100.00 100.00 100.00 100.00
Source: HIPGDAC-ES, Census 1900 - 2021 (INE), GHSL2023A and own elaboration.


Together, our estimates, HIPGDAC-ES, and those of GHSL show notable differences in the few dimensions in which we have compared them. What is not clear is which of the two estimates is more realistic, as everything indicates that GHSL has dispersed the population in excess, with a higher relative number of cells than would be reasonable for the smaller cell size.

5.2 GEOSTAT2021: Census 2021 population grid

Finally, we make a brief comparison of our estimates for 2021, HIPGDAC-ES, with the grid generated by INE from the 2021 census, GEOSTAT2021, which has been generated by bottom-up methods and, in principle, we consider it to be the reference grid14.

The first difference between our estimates and GEOSTAT2021 refers to the number of inhabited cells. While GEOSTAT2021 indicates that we have 115,410 inhabited cells, our estimate gives a value of 143,292 cells –table 2–, i.e. 24% more. This gives us a first indication that HIPGDAC-ES tends to show a greater dispersion of the population than the real one, understanding by this the one offered by GEOSTAT2021.

Table 9 compares the distribution of cells, in relative terms, by population size for 2021 resulting from our estimates, HIPGDAC-ES, with that observed in GEOSTAT2021. This percentage distribution is relatively similar, and discrepancies are only observed at the extremes of the distribution. GEOSTAT2021 shows a lower percentage of very small cells than that observed in HIPGDAC-ES, up to 10 inhabitants, and a higher percentage of larger cells. The discrepancy index, \(100 \times \delta\), between both percentage structures is 2.9, much lower than the one we observed when comparing HIPGDAC-ES with the GHSL, which was 9.8 for the same year.


Table 9. Distribution of cells (%) by population size:
HIPGDAC-ES versus GEOSTAT2021.
Interval HIPGDAC-ES GEOSTAT2021
Up to 10 42.23 39.35
(10, 25] 15.90 15.89
(25, 50] 11.14 11.42
(50, 100] 8.91 9.39
(100, 200] 6.65 6.88
(200, 300] 3.07 3.31
(300, 500] 3.14 3.38
(500, 1000] 3.25 3.53
(1000, 5000] 4.27 4.93
(5000, 10000] 0.80 1.04
(10000, 20000] 0.44 0.62
More than 20000 0.19 0.25
Total 100.00 100.00
Source: HIPGDAC-ES, Census 1900 - 2021 (INE), GEOSTAT2021 (INE)
and own elaboration.


Table 10 compares the distribution of inhabited cells by Autonomous Community and, in this case, the differences are notable, as a result of the fact that HIPGDAC-ES offers a greater number of inhabited cells. In fact, in all regions, except in the Balearic Islands, Navarre, Ceuta and Melilla, GEOSTAT2021 offers a lower percentage of inhabited cells than HIPGDAC-ES, the differences being particularly striking in the Region of Murcia and the Community of Valencia. At the national level, while HIPGDAC-ES indicates that inhabited cells account for 28.1% of the total, GEOSTAT2021 reduces this percentage to 22.6%.


Table 10. Share (%) of inhabited cells by Autonomous Community:
HIPGDAC-ES versus GEOSTAT2021.
Code Region HIPGDAC-ES GEOSTAT2021
01 Andalucía 30.0 21.6
02 Aragón 11.1 7.9
03 Principado de Asturias 51.8 48.9
04 Illes Balears 54.6 64.9
05 Canarias 47.7 46.5
06 Cantabria 42.8 41.1
07 Castilla y León 15.3 13.0
08 Castilla-La Mancha 14.7 7.8
09 Cataluña 47.2 40.2
10 Comunidad Valenciana 47.5 36.2
11 Extremadura 17.8 9.6
12 Galicia 64.8 62.5
13 Comunidad de Madrid 41.2 35.3
14 Región de Murcia 50.0 35.9
15 Comunidad Foral de Navarra 17.9 18.8
16 País Vasco 54.5 50.1
17 La Rioja 17.5 13.1
18 Ceuta 64.1 69.2
19 Melilla 47.2 50.0
Total 28.1 22.6
Source: HIPGDAC-ES, Census 1900 - 2021 (INE), GEOSTAT2021 (INE)
and own elaboration.


Table 11 compares the percentage distribution in terms of the altimetric cuts and shows a high level of agreement. The discrepancy index, in percentage terms, \(100 \times \delta\), is only 0.28.


Table 11. Population share (%) by altitude (m):
HIPGDAC-ES versus GEOSTAT2021.
Altitude HIPGDAC-ES GEOSTAT2021
Up to 200m 53.0 53.3
(200, 600] 22.0 21.9
(600, 1000] 23.7 23.6
More than 1000m 1.3 1.2
Total 100.0 100.0
Source: HIPGDAC-ES, Census 1900 - 2021 (INE), GEOSTAT2021 (INE)
and own elaboration.


Finally, table 12 compares the percentage distribution in terms of distance to the coast. In this case the percentages in the various cutoffs are virtually identical. The discrepancy index, in percentage terms, \(100 \times \delta\), is only 0.12.


Table 12. Population share (%) by distance to the coast (km):
HIPGDAC-ES versus GEOSTAT2021.
Distance HIPGDAC-ES GEOSTAT2021
Up to 10km 39.5 39.4
(10, 50] 18.6 18.7
(50, 100] 10.7 10.7
(100, 200] 10.7 10.8
More than 200km 20.5 20.5
Total 100.0 100.0
Source: HIPGDAC-ES, Census 1900 - 2021 (INE), GEOSTAT2021 (INE)
and own elaboration.


Overall, the results show a high degree of agreement in relative terms, although significant discrepancies remain in absolute terms, i.e. in terms of inhabited cells.

It is illustrative to calculate the confusion matrix –crosstabulation– between the two grids for 2021. This matrix is given in both, absolute and relative terms, in table 13. Of the inhabited cells in GEOSTAT2021 only 8.2% –9,458 cells– are not inhabited in HIPGDAC-ES. Thus, almost all the inhabited cells in the INE census grid are also inhabited in our estimates. On the contrary, of the inhabited cells in HIPGDAC-ES, there are 26.1% of cells –37,505 cells– that are not inhabited in GEOSTAT2021.


Table 13. Cross tabulation of cells: HIPGDAC-ES versus GEOSTAT2021.
GEOSTAT2021
Cells
Percentages (%)
YES NO Total YES NO Total
HIPGDAC-ES
YES 105,952 37,505 143,457 20.72 7.34 28.06
NO 9,458 358,379 367,837 1.85 70.09 71.94
Total 115,410 395,884 511,294 22.57 77.43 100.00
Source: HIPGDAC-ES, Census 2021 (INE), GEOSTAT2021 (INE) and own elaboration.


To informally examine the discrepancies in absolute terms, it is illustrative to compare both grids, HIPGDAC-ES and GEOSTAT2021, on a map. Maps 1 and 2 give this visual impression for Murcia and Castellón, where, according to table 10, the discrepancies are notable.

In the case of Murcia, map 2, the greatest dispersion of the population –in the sense of a greater number of inhabited cells- is evident, especially in the interior and in a more dispersed rural population.

Mapa 2: HIPGDAC-ES versus GEOSTAT2021 - Murcia: Inhabited cells.

Source: HIPGDAC-ES, Census 2021 (INE), GEOSTAT2021 (INE) and own elaboration.

In the case of Castellón, map 3, we observe a similar effect, especially in the interior of the province, although somewhat less pronounced. It seems reasonable to posit that the existence of buildings classified as residential in the cadastre allows the population distribution algorithm to assign population to them, although in practice they do not house any resident population.

It is possible to offer a quantitative measure of the discrepancy between both grids by a simple modification of the \(\delta\) discrepancy statistic. In this case, both grids add the population of the 2021 census, so that to narrow the index to the interval [0, 1] it is enough to rescale it by the population:

\[\delta ' = \frac{{\Sigma_c}\left| {{P_c} - P_c^{ref}} \right|}{2.{\Sigma_c}{P_c}} \tag{2}\]

where \(c\) indexes the grid cells.

This statistic, \(\delta '\), ranges between 0, if both grids are identical, and 1, in the case of maximum discrepancy, when there is no overlap between the inhabited cells in both population grids. Thus, and assuming that GEOSTAT2021 is the true distribution of the population, \(P_c^{ref}\) in (2), \(100 \times \delta '\) can be interpreted as the percentage of population that we place incorrectly on the territory, where the accuracy or the inaccuracy in the location is determined by the grid resolution.

The value of \(\delta '\), in percentage terms, \(100 \times \delta '\), for HIPGDAC-ES and GEOSTAT2021 is 11.7. Which indicates that our disaggregation algorithm places, in 2021, about 12% of the population over the territory incorrectly.

Mapa 3: HIPGDAC-ES versus GEOSTAT2021 - Castellón: Inhabited cells.

Source: HIPGDAC-ES, Census 2021 (INE), GEOSTAT2021 (INE) and own elaboration.

6. Concluding comments

This paper presents a general methodology to generate census population grids from conveniently homogenised historical cadastral information (Uhl et al 2023) and homogeneous population series at municipal level (Goerlich, Ruiz, Chorén and Albert 2015) from 1900 to the last available census. The generated information is available as raster layers in resolutions 100m x 100m and 1km x 1km, and, in the latter case, the grids are also available in vector format.

Although validation of the results is complicated by the lack of historical data, an examination of the trends shown indicates that the results appear reasonable. Comparison with the GHSL population grids indicates that our estimates appear more accurate than in these grids, which tend to over-disperse the population, and generate a very high number of inhabited cells. A comparison with the INE census 2021 grid, GEOSTAT2021, yields encouraging results, as they are relatively coincident in many dimensions, although it is clear that our estimates tend to over-disperse the population, especially in rural areas. In fact, this is a well-known bias of the dasymetric methods of spatial disaggregation (Gallego 2010), given the difficulty of limiting the allocation of population to residential areas that are not the main residence for most people. Overall, although the estimates seem far from perfect, they are an improvement on the population distribution from municipal data, where the population is either concentrated in one point –a cell– or spread evenly throughout the municipality.

It should be noted that this work offers initial estimates of grids of census population from 1900 to the present, but that it is an initial estimate –version 1– that must be submitted to validation by experimentation of the results obtained and, if possible, enhanced with additional available information. Therefore, it should be borne in mind that this is work in progress, that can be improved in the future, and on which any comment is welcome. The development of alternative methods of generating historical population grids, based on the geocoding of gazetteer settlements (Diez Minguela, Goerlich and Tirado Fabregat 2024), can shed light on the quality of these estimates.

Referencias


  1. This version of the database represents a slight modification with respect to the initial one –version 0–, that appears in the Working Paper of the Valencian Institute of Economic Research (Ivie).

    The modification refers to the fact that, in order to avoid border effects of administrative boundaries, the search for support for the population of municipalities in the Basque Country is limited to those layers that are known to exist in those Cadastres. Consequently, only a few municipalities in the Basque Country are affected.↩︎

  2. The version used in this work is version 2, which corresponds to the final published work (Uhl et al 2023), and is substantially improved from the initial version. The database is available in figshare.↩︎

  3. European Terrestial Reference System 1989 (ETRS89) - Lambert Azimutal Equal Area (LAEA).↩︎

  4. In fact, the variables related to residential buildings only, RES_BIA and RES_BUFA, were included in the second version of the HISDAC-ES database, which is the one finally published, with the intention of facilitating the integration with historical population data and facilitating its spatial disaggregation (Uhl et al 2023, p.- 4724). Its periodicity, decennial in line with the cesus, and its CRS, ETRS89-LAEA, are lower than the rest of the layers, since they are intended for statistical use. Therefore, our work represents a natural extension of HISDAC-ES.↩︎

  5. The consultation was carried out through Luis Julián Santos Pérez, Chief of Service of the Directorate General of Cadastre, and specialist in the historical cartography of Cadastre (Santos 2019).↩︎

  6. The ESPAREL project has been financed by the BBVA Foundation within the Program of ‘Aid to Scientific Research Teams in Digital Humanities’, call 2019.↩︎

  7. These municipalities are Badia del Vallès (08904), Vegaviana (10902), Alagón del Río (10903) and Vencillón (22909), and its population only exceeds 1,000 inhabitants for Badía del Vallès (08904) in 1950. In the other years the municipalities for which we lack support to allocate the population are extremely small. For example, in 1900 these populations ranged between 276 inhabitants for Badía del Vallès (08904) and 5 inhabitants for Alagón del Río (10903).↩︎

  8. These results are not invariant to the geographical scale at which we round to integers, so the numbers would be different if we rounded at the 100m x 100m cell level.↩︎

  9. Given the population increase over the period 1900 - 2021 all results are shown in relative terms.↩︎

  10. The rest of the trends, the displacement towards urban municipalities and provincial capitals (Goerlich and Mollá 2021), cannot be examined on the basis of the generated information, since the origin of it is municipal data.↩︎

  11. Distances are measured in a straight line from the cell to the coastline of the layer of administrative contours of IGN (Goerlich and Perez 2021). Calculations were made on the grid projection, ETRS89-LAEA.↩︎

  12. Of course, grids for the years 2025 and 2030 are projections.↩︎

  13. The information is also provided in geographic coordinates, WGS84 reference system, with resolutions of 3 and 30 arcseconds.↩︎

  14. The reasons why we did not make the same comparison with the grid generated by INE from the 2011 census, GEOSTAT2011, is because its production method is not comparable with GEOSTAT2021 and generated some undesirable results (Goerlich 2024, section 5).↩︎