Making the Map

By: David Gilman Romano, Ph.D., Nicholas L. Stapp, Ph.D. & Andrew B. Gallia, Ph.D.

The attached maps utilize a wide range of published information to create an entirely new, comprehensive, and accurate period plan of the Augustan-period city. In addition to indicating the size, location, and orientation of the various structures, roads, and water systems of the city, we have also attempted to present here a detailed, reasoned contour map of the city’s historical topography at the time of Augustus. Throughout, the goal has been to produce maps that are both attractive and easy to read, while maintaining a high degree of accuracy.

Cartographic resources

The maps have been created through the use of the facilities of the Corinth Computer Project Research Laboratory at the University of Pennsylvania Museum. The principal software programs used in the production of these maps include: Autocad and the extension module CAD Overlay, which were used for data collection and drafting; Surfer, which was used in creating the model of the historical elevations; and Adobe Photoshop and Adobe Illustrator, which were used to improve the aesthetic appeal of the finished maps (see Appendix A below).

The primary basis for our map is a set of 1 : 10 000 maps from the Carta tecnica regionale of the Regione Lazio, based on aerial photographs and a topographic survey carried out in 1990.1 Our study area, a region of c.20 km2 surrounding the centro storico, is split between four different maps in this series, so we began by bringing together the relevant sections of these maps.2 Each of these four maps was scanned commercially at a scale of 1 : 1. Using CAD Overlay, the digital copies (rasters) of these maps were scaled, rotated, and then rectified to a static grid of 500 m. squares over their entire area.3 This process made the digital version of the maps more accurate than the printed form in which we had received them. Subsequently the four maps were merged and cropped to create one seamless raster image. Once rectified and merged, this digitized 1 : 10 000 scale map became the baseline for accuracy to which all other cartographic information was compared; ultimately all of the archaeological information presented in our maps was positioned with reference to this ‘base map’ of the modern city.

Another important source of modern cartographic information used in creating these maps was the Atlante di Roma, an aerial photographic atlas of the centro storico reprinted at a scale of 1 : 1000. The detailed information provided in this publication proved useful for the interpretation of some areas of the Carta tecnica regionale. It was also a great source of information for the creation of the topographical model (see below). In addition, the street map included in the Roma guide of Touring Club Italia was used for the quick location and identification of street names, as well as for resolving other issues raised by the Carta tecnica regionale.

Archaeological maps and plans

With the raster version of the Carta tecnica regionale as our base map, we began by assessing the accuracy of Scagnetti’s map of 1979.4 This was scanned at a high resolution and, using Autocad and the CAD Overlay module, was matched against our rectified base map of the modern city. When we attempted to fit the two images together, we found that it was not possible to fit the Scagnetti map (or even portions of it) to our base map without creating gross distortions of scale. Therefore it could not be used as a primary source of topographical information in the creation of our map.

A similar, though less troublesome, problem was encountered with the maps of the Carta archaeologica di Roma. These annotated large-scale maps (1 : 2500) provide the locations and detailed descriptions of archaeological remains in the N section of the city. The scanned versions of these maps also could not be fitted to our base map over their full extents without excessive distortion. Nevertheless, it did prove possible to take some smaller (500 x 500 m) sections of these maps and match them against the base map so that they fitted within an acceptable tolerance of +/-10 m.

An extremely useful map was the 1 : 2000 tourist map of the Forum.5 When scanned and fitted against the digital base map, this map was found to be highly accurate. Because of its larger scale, this map was able to provide much information that was not available in the 1 : 10 000 scale of our base map. It was helpful in providing a detailed plan of the Capitoline and Palatine hills and especially for the archaeological zone of the Forum Romanum, which is depicted only schematically in the modern Carta tecnica regionale.

Another useful map was the Carta del centro storico di Roma.6 It provides detailed plans of the archaeological remains together with proposed reconstructed plans of ancient structures overlaid against the modern street plan. When these maps were scanned at high resolution and fitted against our base map, they were found to match up quite well.7 The larger scale of these maps (1 : 1000) combined with their great accuracy make them an important resource, especially for locating the actual location of preserved remains, although their reconstructions of the plans of ancient structures may be disputed at times.8

Still indispensable to any cartographer of ancient Rome are the plates of Lanciani’s Forma Urbis Romae, published in 1893 at a scale of 1 : 1000. While serious questions have been raised about Lanciani’s methodology, his plans still provide the most detailed map of ancient remains available for the city as a whole.9 In our project the Forma Urbis Romae was commonly used for mapping the course of roads and aqueducts. Although some distortions have been noted, the overall accuracy in the scale of these plans is quite exceptional. The primary difficulty encountered stems from the transformations that have taken place since the 19th c.; especially in the S and E sections of our study area, modern development has made it difficult to position some of Lanciani’s plates.

Mapping the Augustan city

The bulk of the work went towards the process of taking information from a wide range of sources and combining it with the data contained in the above-mentioned maps. This meant using the CAD software to draft individual plans for each of the buildings, roads, waterways, and other features that have been included on the map, and to place these features in relationship to one another on the basis of their location within the modern city plan. While some information was taken directly from the above-mentioned archaeological maps, the more usual procedure was to start from a separately-published actual-state or restored archaeological plan of each structure. For every item that has been mapped, the author of the corresponding entry submitted a copy of the best available plan of the structure, along with as much information as possible about its size, orientation, and precise location. These drawings were scanned and then opened in Autocad using the CAD Overlay module. Next, they were scaled, rotated, and fitted into their proper location in relationship to the rectified modern ‘base map’ of the Carta tecnica regionale. Once this raster information was in place, each structure was drawn according to the visual conventions developed for our map, following the interpretation of the author.

The procedure for mapping larger structural systems, such as roads, aqueducts, and sewers, was sometimes different from that used for individual monuments. The most useful plans of these features are typically ones that show the position of the ancient remains in relationship to a modern street plan. Because of distortions in some modern maps, it was not always possible to fit these plans to our base map satisfactorily. When this occurred, we divided the smaller-scale plans into sections, each of which could then be positioned with greater accuracy. If we could rely on the basic topological accuracy of these plans, this process of sectioning them should allow us to recover a more accurate outline of a road’s or aqueduct’s course from a less accurate plan. Whenever possible, extant sections of these larger networks were taken from more detailed (and more accurate) large-scale maps and excavation plans.

Another set of methodological problems surrounds the use of evidence from the Severan Marble Plan. For some of the buildings and roadways on our map, this fragmentary 3rd-c. A.D. city plan is the sole source of detailed information. To depict these structures, we relied on maps that locate the Marble Plan fragments in relationship to modern Rome, such as those presented by Rodríguez Almeida.10 In general, we sought to retain the integrity of a structural footprint inscribed on the Marble Plan. Some instances, however, required minor modifications to information on the Marble Plan; using the drafting capabilities of the computer, minor inaccuracies by the stone mason could be corrected, such as straightening slightly bent lines or making the two sides of a portico truly parallel.

In mapping the physical remains of buildings, sewers, or walls, we applied consistent standards to ensure a high degree of accuracy. In addition to mapping the restored ground-plans of various structures, we have indicated by a cross "†" the position of those remains for which reconstructions were not possible.

Our map developed gradually, as plans and other information were submitted and then converted into drawn objects using Autocad. The process has been a collaborative one, with each of the contributors working closely with the mapmakers to ensure that the cartographic representation of each structure corresponded as closely as possible to his or her written text.

We also present a great deal of speculative material in our map. For monuments lacking extant remains and a certain location, we could not follow the rigorous drafting procedures outlined above. Many structures could only be positioned relative to other buildings; for these, each author presented sketches indicating a hypothetical location. The authors also helped define the ‘area shadings’, which present a general sense of the urban mass and development of the ancient city. While the green-shaded areas indicating garden-estates (horti) were drawn in consultation with the various authors responsible for those entries, the shape and extent of the built-up areas and defined urban spaces were developed over the course of this project, and represent the consensus of the authors. The specific course and width of the Tiber (which varied in antiquity) have been arbitrarily defined; our depiction is based roughly on its modern course, with some modifications.11

Modeling the reconstructed topography of Augustan Rome

The contour map of the ancient landscape was produced using a computer modeling process developed specifically for this project. Work on the contour model began simultaneously with the creation of the archaeological plan and proceeded in conjunction with it. We began by creating an accurate model of the city’s modern topography, based upon the same 1 : 10 000 Carta tecnica regionale base map. We then made reasoned changes to this model to reverse the effects time has had on the topography. The methodology aims at making the reasoning that went into creating our model as explicit as possible. Each hypothetical contour line can be justified with reference to the modern topography and to the modifications that our understanding of the archaeological and geological situation suggested should be made.

Step 1: Modeling the modern city

The first step in this process was the creation of a computer model of Rome’s modern topography.12 To model the shape of the slope of the hills accurately, it was necessary to have as much data as possible. Our two main sources of information were the rectified digital base map of the Carta tecnica regionale, and more detailed plans in the Atlante di Roma, both of which provide elevations for various points taken from aerial surveys. These two resources were supplemented with information drawn from a number of more specialized studies.13 Using Autocad, this elevation information was digitized as a series of Z-coordinate values assigned to points plotted against the two-dimensional space of our base map in the XY plane. Altogether, we digitized some 2,476 data points, whose location could be exported as a set of XYZ coordinates for use by Surfer, our topographic modeling software package.

In order more accurately to represent the natural topography, it was necessary to exclude those data points the elevation of which reflects human intervention after the Augustan period. These include high elevation points taken on roofs, high sidewalks, the tops of fountains or other monuments. We also excluded data for places where the ground has been raised or lowered artificially, such as bastions, terraces, excavations, embankments, and the ramps leading up to bridges. Elevation points on Monte Testaccio were also ignored.

The Tiber River was given special treatment in the creation of our model. The modern embankment creates a steep drop-off, a feature incongruent with the terraced embankments of antiquity.14 The course of the river has also changed since Augustus. In order to prevent the distortions that the modern embankment would create in the vicinity of the river, we excluded all of the elevation data given for points inside the embankment walls.

From the 2,476 data points collected, we created a digital model of the modern landscape of the city using Surfer. This software package applies mathematical algorithms to unevenly spaced XYZ data to generate an elevation model of Z-coordinate values across a grid of evenly distributed points in the XY plane.15 This matrix of values can then be combined mathematically with other matrices (representing areas of alluviation, etc.) to introduce hypothetical changes into the topography of the study area. Surfer also provides a function to create conventional contour maps from these elevation grid models. The contour map of our model of the modern topography is shown in fig. 3.

Fig. 3. Rome, contour map of the modern surface topography (5 m intervals). Scale 1 : 30 000.

Step 2: Introducing changes

The next step was to establish what changes needed to be imposed upon the model of the modern city to recreate the topography of the Augustan period. The first set of changes were ones to counteract the impact made by two millennia of alluviation from the Tiber. We began by defining the extent of the city that has been altered by alluvial deposits, following the plans presented in the recent geological study of Rome’s centro storico.16 This map was scanned at high resolution and used to define areas where the ground level should be lowered.17 The information provided by this large-scale geological map was supplemented with more detailed information available in recent studies of the Forum basin, the Capitoline, Palatine, and the Velabrum.18

Fig. 4. Rome, areas of constant alluviation. Scale 1 : 30 000.

Because the amount of sedimentation that has taken place since the Augustan period varies from place to place, we decided to divide the broader floodplain into smaller sections, each of which could be correlated with what is known about Augustan elevations in that area of the city. Bringing together information about the difference between modern and Augustan levels in the Tiber floodplain, we were able to produce a rough map of areas of constant alluviation (fig. 4). This information can also be represented in the following table:

Table 1: Rome, known Augustan-era grounds levels

Location Approx. distance below modern level Source
N and NE Campus Martius 6 m Rakob 1987, fig. 7
Central Campus Martius 7 m Rakob 1987, fig. 7; Manacorda & Zanini 1989, 25
Trans Tiberim 5 m Lanciani, FUR pl. 20
Forum Boarium 5 m cf. Coarelli 1988, 38 fig. 4
Imperial fora 7 m Einaudi 1988, fig. 1; Amici 1991, fig. 15
Forum Romanum 8 m Einaudi 1988, fig. 1; Amici 1991, fig. 15
Circus Maximus 7 m Ciancio Rossetto 1987
Monte Testaccio 6 m Rodríguez Almeida 1984

We have also identified other areas where the level of the ground surface has been raised due to human intervention. This includes Montecitorio and Monte Giordano, the small hillocks in the Campus Martius which formed above the accumulation of archaeological remains.19 Another major change appears on the W face of the Quirinal hill, where soundings revealed that the gardens of Palazzo Quirinale stand atop an infilled valley whose floor lies as much as 23 m. below the modern level.20 These areas of artificial elevation were mapped with Autocad and used to alter our model of the modern topography.

There are also areas that are much lower now than during the Augustan period. The process of accounting for lost elevation presents a different set of challenges. In general, we have been unable to account for the effects of erosion and landslides on the shape of Rome’s hills. There is not enough reliable data upon which to base a reconstruction of features that have been lost in this way. Therefore some of the hills may have shallower sides and smaller summits than they actually possessed in the Augustan period.

Only in two cases have we attempted to reconstruct features that have been lost from Rome’s topography. The first is the low saddle that once ran between the Quirinal and the Arx before being cut away to make room for the Forum of Trajan.21 We have also tried to reconstruct the rise of the Velia, which has been obscured by the Hadrianic temple of Venus and Roma.22 Hypothetical contours of these features were drawn and used to interpret how the modern topography needed to be modified.

In other areas of the city, especially along the tops of the hills, today’s ground level is still very close to what it was at the time of Augustus. We have taken those areas on the geologic map where the dominant soil type is tufa to correspond to zones where the elevation has remained essentially constant. These were drawn as areas requiring no change. The space between these areas and the areas in the floodplain, along the slopes of the hills, was left empty.

The above information was entered into a cumulative map of changes to be applied to the model of modern topography, showing where the ground level should be reduced, increased, or left constant. We then generated a series of points across this map, assigning each a Z-coordinate value that was equal to the amount that the elevation was supposed to be changed in each region. For example, all of the points in the region of the N Campus Martius were assigned a Z-coordinate value of -6 m, indicating that the ground level in that region should be reduced by 6 m. These points were then exported to the Surfer software package and used to create a topographic grid model of exactly the same size as the model of the city’s modern topography. In those areas where the amount of change was not set (i.e., along the slopes of the hills), the computer calculated a quantity based upon the surrounding values. The model thus created provides that the greatest amount of elevation will be removed in the floodplains; that amount becomes gradually less as one moves up the sides of a hill, until reaching the top, where the amount of change is zero. A map representing this model is shown in fig. 5.

Fig. 5. Rome, model of all changes applied to the modern topography. Scale 1 : 30 000.

Once this model was created, it was combined with the model of the modern topography. The matrix of values for modern elevations was added to the matrix of values for the changes to produce a third matrix representing the elevations across a grid for the topography of Augustan Rome. From this, we produced the map of contour lines representing the surface topo-graphy of Augustan Rome (fig. 6; to compare the ancient and modern topographies, juxtapose figures 6 and 3).

Fig. 6. Rome, contour map of the Augustan surface topography (5 m intervals). Scale 1 : 30 000.

Design of the map

When the main work was nearing completion, we approached Mark Davison, a professional graphic designer, to help improve the overall legibility and aesthetic impact of the map. A graduate of the University of Pennsylvania’s Department of Landscape Architecture who had previously worked for the Corinth Computer Project, he added a new perspective to our work. First, the final schematic version of the map, which combined both the archaeological and topographical data with the modern street plan, was converted from Autocad to a format that could be used with graphic design software that would allow a much broader range of colors and more sophisticated visual effects. This involved manipulating the files using Adobe Illustrator to fix the appropriate scales for the two maps. From here, the remainder of the work, including setting the color palette, applying various visual effects, and plotting the final layout of the map, was done in Adobe Photoshop.

Archaeological Mapping Lab
University of Arizona

Appendix A: Computer and software resources


  • 2 Pentium class PC’s running Windows 98 and NT
  • Large-format color plotter
  • Small-format scanner
  • Large-format scanner (outsourced)


  • Autocad R14 and 2000
  • CAD Overlay 8 and 2000
  • Surfer v.6
  • Adobe Photoshop 5.5
  • Adobe Illustrator 9.0

Appendix B: Map resources

  • (1 : 1000) R. Lanciani, Forma Urbis Romae (Rome 1893-1901; repr. 1990 at a scale of 1 : 2000).
  • (1 : 1000) E. Guidoni (ed.), Carta del centro storico di Roma (Rome 1985-92) nos. 29, 31, 38, 39, 40, and 49.
  • (1 : 1000) Comune di Roma, Atlante di Roma: la forma del centro storico in scala 1 : 1000 nel fotopiano e nel carta numerica (2nd edn., Venice 1991).
  • (1 : 2000) Soprintendenza Archeologica di Roma, in collaboration with the Tourist Office of Rome, Rome. Archaeological centre (Rome 1985).
  • (1 : 2500) Commissione per la Carta archeologica d’Italia, Carta archeologica di Roma I-III (Florence 1962-77).
  • (1 : 5555, stated as 1 : 5000) F. Scagnetti and G. Grande, Roma urbs imperatorum aetate (Rome 1979; 2nd edn. 1986).
  • (1 : 10 000) Regione Lazio, Carta tecnica regionale (Rome 1990) nos. 374060, 374070, 374100, 374110.
  • (1 : 10 000) A. Corazza and P. Lombardi, “Carta idrogeologica del centro storico di Roma,” in R. Funiciello (ed.), La geologia di Roma. Il centro storico (Rome 1995) pl. 1.
  • (1 : 10 000) F. Marra and C. Rosa, “Carta geologica del centro storico di Roma in scala 1 : 10.000,” in R. Funiciello (ed.), La geologia di Roma. Il centro storico (Rome 1995) pl. 9.
  • (1 : 12 500) A. Agnati (ed.), Roma. Guida d’Italia del Touring Club Italiano (8th edn., Milan 1993) inset map.
  1. W. Alvarez et al., “Quaternary fluvial-volcanic stratigraphy and geochronology of the Capitoline Hill in Rome,” Geology 24 (1996) 751-54.
  2. M. Amanti, P. Cara, and M. Pecci, “Modello digitale della superficie reale del centro storico della città di Roma,” in Funiciello 1995, 309-21.
  3. C.M. Amici, Il Foro di Cesare (Florence 1991).
  4. A.J. Ammerman, “On the origins of the Forum Romanum,” AJA 94 (1990) 627-45.
  5. A.J. Ammerman, “Morfologia della valle fra Palatino e Velia,” BA 16-18 (1992) 107-11.
  6. A.J. Ammerman, “Environmental archaeology in the Velabrum, Rome: interim report,” JRA 11 (1998) 213-23.
  7. A.J. Ammerman and N. Terrenato, “Nuove osservazioni sul colle Capitolino,” BullCom? 97 (1996) 35-46.
  8. J.C. Anderson, The historical topography of the Imperial Fora (Brussels 1984).
  9. G. Angelis d’Ossat, “L’antica topografia del colle Quirinale,” BullCom? 66 (1938) 5-15.
  10. P. Carafa, “Il tempio di Quirino: considerazioni sulla topografia arcaica del Quirinale,” ArchCl? 45 (1993) 124-28.
  11. F. Coarelli, Il foro Boario dalle origini alla fine della repubblica (Rome 1988).
  12. H. de Wit, “Lineamenti dell’ambiente fisico: alcune considerazioni sui sondaggi nell’ospedale militare e nell’area di Piazza Celimontana” in C. Pavolini (ed.), Caput Africae I (Rome 1993) 213-18.
  13. R. Einaudi, “Excavations in the Roman status quo,” Places 5 (1988) 36.
  14. R. Funiciello (ed.), La geologia di Roma: il centro storico (Rome 1995).
  15. M. Lenoir, “Vigna Barberini,” BullCom? 91 (1988) 507-13.
  16. D. Manacorda and E. Zanini, “The first millennium A.D. in Rome: from the Porticus Minucia to the Via delle Botteghe Oscure,” in K. Randsborg (ed.), The birth of Europe: archaeology and social development in the first millennium A.D. (ARID suppl. 16, 1989) 25-32.
  17. F. Marra and C. Rosa, “Stratigrafia e assetto geologico dell’area romana,” in Funiciello 1995, 49-118.
  18. C. Panella (ed.), Meta Sudans I (Rome 1996).
  19. F. Rakob, “Die Urbanisierung des nördliches Marsfeldes,” in L’Urbs 1987, 687-712.
  20. D.W. Reynolds, Forma Urbis Romae: the Severan marble plan and the urban form of ancient Rome (Ph.D. diss., Univ. of Michigan 1996).
  21. E. Rodríguez Almeida, Forma urbis marmorea: aggiornamento generale 1980 (Rome 1981).
  22. E. Rodríguez Almeida, Il Monte Testaccio: ambiente, storia, materiali (Rome 1984).
  23. M. Sediari, “Carta archeologica del Campo Marzio,” in C. D’Onofrio (ed.), Via del Corso: una strada lunga 2000 anni (exh. cat., Rome 1999) 172-75.
  24. E. Tortorici, “Alcune osservazioni sulla tavola 8 della ‘Forma Urbis’ del Lanciani,” QuadTop? 10 (1988) 1-16.
  25. U. Ventriglia, La geologia della città di Roma (Rome 1971).