The Making of My Middle‑earth Maps

As a young lad, the lands of Middle‑earth held a particular sense of longing for me, and they still do, although I am now a slightly older one. For me, the landscapes in which the tales take place are one of the great beauties of the legendarium. They invite exploration. And great landscapes invite great maps.

The maps included in the books did what they needed to do. They showed the important places, distances and routes, and allowed the reader to keep his bearings while following the characters on their journeys. Yet beside the extraordinary richness of Tolkien’s writing, I always wanted to see more of the land itself. Most other maps of Middle‑earth still looked like fantasy maps to me. I wanted something that felt closer to a real-world map, as though the land had actually been surveyed.

My first attempt began in 2014 and was largely drawing-based. I was inspired in particular by the large National Geographic maps, which could contain an extraordinary amount of information while remaining clear and believable. However, I soon realised that manual drawing alone would not give me the realism I was looking for, nor the intricate detail that makes a map feel convincing. It was also an extremely slow way of working.

I eventually put that first version aside. It had nevertheless taught me a great deal, in particular that the map needed a more solid technical foundation. If I wanted it to feel like an actual surveyed map rather than an illustrated interpretation of one, I would have to begin again and approach the work in a different way.

If I wanted to make something that felt like a real-world map, it seemed sensible to begin by understanding how such maps are actually made. I started looking more closely at topographic and terrain maps, elevation data and the methods used to describe the form of a landscape rather than merely illustrate its visible features.

This changed the direction of the work entirely. Instead of drawing the finished terrain directly, I began developing digital elevation models from which the landscape could be built. This gave the map the physical structure required for everything that would later be added. The terrain itself could now guide how the landscape developed, rather than every visible feature having to be drawn separately.

Over 2022 and 2023, I worked on developing this new approach and building the first terrain-based Middle‑earth map. By then I also had considerably more technical experience than when I made the first attempt. I had worked in Photoshop for roughly twenty years, and over time had become much better at complex image editing and handling very large files. The maps were built through this combination of terrain data and manual image work, without the use of AI or machine learning.

Even so, much of the project was new to me. I had to learn new software, understand how geographic terrain data behaved, and find ways to adapt it to the particular needs of the map. I wanted to use the methods and underlying principles of real terrain mapping, while remaining as close to Tolkien’s geography as I could.

Before developing the terrain, I spent a great deal of time studying the available maps and drawings. Christopher Tolkien’s published maps formed the main geographic framework, but I also looked extensively at Tolkien’s own sketches and drafts. I was interested not only in what he drew, but in what he later changed. Repeated alterations to certain parts of the map could give some indication of which areas were particularly important to him.

I also compared Tolkien’s annotated map, maps prepared by Stephen Raw for later editions, and the revised edition of Karen Wynn Fonstad’s atlas. Each source was useful in a different way, although they do not always agree. Positions, distances and forms can vary, and some features are shown with more certainty in one source than in another.

Where a clear published solution existed, I generally followed it. Where the references differed, I compared them with the surrounding geography and tried to choose the interpretation that seemed most consistent with the whole. This was not an attempt to correct Tolkien or produce a definitive version. I wanted to understand the decisions behind the maps as well as I could, and then remain as close to them as the work allowed.

For me, serious Tolkien-inspired work begins with respect for the source material. There is room to interpret what is not fully shown, yet this is not my world to correct.

The basis of the terrain is a digital elevation model, usually shortened to DEM. In simple terms, this is an image in which the brightness of each point represents its height. It is not yet a finished map, but a representation of the land from which the visible terrain can later be developed.

I worked with 32-bit floating-point elevation files and kept the data in that form throughout the DEM workflow. This retained the available height information and avoided damage from unnecessary conversion or compression. The completed DEMs were built from hundreds of separate elevation images and terrain sections. These had to be adjusted, blended and often reshaped by hand before they could function as one larger landscape. Much of the terrain modelling was therefore manual, even though it began with geographic elevation data.

Real terrain data was useful because it contained the complexity I had been unable to achieve through drawing alone. However, an existing landscape could not simply be copied into Middle‑earth. Each section had to be judged for its scale, direction and character, and then adapted to Tolkien’s geography. Much of the work lay in deciding what could remain, what needed to be changed, and how separate pieces could be made to belong together.

This was by far the most time-consuming part of the map. The DEM had to support the known geography while also providing enough detail in the areas Tolkien’s maps do not fully describe. It was revised continually, as later stages often revealed forms in the terrain that were difficult to judge from the black‑and-white elevation model alone.

Once the DEM was sufficiently developed, I could begin turning the elevation data into visible terrain. This was done through hillshading, also known as shaded relief, where a virtual light is used to reveal the shape of the land through light and shadow.

A single hillshade was not enough. One lighting direction could describe certain forms well while obscuring others, and stronger settings could easily make the terrain appear too harsh or artificial. I therefore produced many different hillshading passes and combined them so that the whole formed a reasonably lively and three-dimensional landscape.

The purpose was not simply to make the mountains stand out. The shading also had to remain readable across flatter regions and allow smaller changes in elevation to remain visible without overwhelming the map. Much of the work involved balancing these different scales of terrain so that the landscape could be understood both from a distance and under closer inspection.

The hillshading continued to change as the map developed. Colour and texture could alter how strongly a slope appeared, so each pass had to be judged as part of the complete terrain rather than as a finished layer on its own.

Colour was developed in relation to the elevation model rather than applied as a flat layer over the finished terrain. It added another level of visual interest to the map and helped reveal smaller forms and textures that could not be described through hillshading alone.

I worked with several colour passes, each serving a different purpose. Some introduced broad regional differences, while others added more local variation within the terrain. The aim was to avoid large areas feeling uniform, without making the colour so strong that it competed with the geography.

The treatment also had to reflect the character of the different regions. Helcaraxë and Forodwaith required a colder and more subdued appearance, while Haradwaith called for warmer tones. These changes were kept gradual, as the regions still had to belong to the same landscape rather than appear as separately coloured areas.

The colours also had to work in print. A combination that appeared clear on screen could become too dark, too strong or too similar once printed. I therefore returned to the colour work repeatedly, making small adjustments until it felt like a natural part of the whole.

Once the terrain and colour were established, I began adding the surface layers that would give more interest and life to the map. Regional variation and smaller details helped prevent large areas from feeling too uniform.

The texture had to belong to the terrain rather than appear as a separate layer placed over it. It was adjusted according to scale and location, and often kept deliberately subtle. Too much detail could make the map difficult to read, while too little left the landscape feeling bare.

Forests presented a particular problem. On many maps they appear as large, clearly bounded shapes, which is useful for showing their general position but can quickly feel unnatural when placed over realistic terrain. I kept the main forests relatively muted and added looser patches of woodland around them where this seemed natural. I could be somewhat freer here, as smaller areas of woodland would not alter the geography or contradict the stories.

I treated rivers with considerably more caution. Additional rivers would have made parts of the map appear more realistic, yet an important river would quite likely have been mentioned in the texts. I therefore retained the known rivers and generally avoided inventing new systems, apart from one or two minor exceptions in less important regions.

Considerable work had already gone into forming the rivers within the DEM itself. They do not simply sit as lines over an unrelated landscape, but follow channels that were built into the terrain from an early stage. This helped the water belong more naturally to the surrounding land.

The labels and symbols were important in giving the maps a more modern, real-world appearance. I was again influenced by large National Geographic maps, where the information follows a clear visual system and can be understood without drawing too much attention away from the landscape.

The lettering required a consistent hierarchy. The size, weight and style of a name had to reflect what it referred to and its importance within the map. Once this system was established, it became easier to distinguish larger regions, settlements and other geographic features at a glance.

I kept the symbology as simple and clean as possible. Roads, settlements and other marked features needed to be recognisable, yet the symbols could not become decorative elements in their own right. Their purpose was to provide information clearly and remain subordinate to the terrain.

Everything was judged at the intended print size. Labels and symbols had to remain legible without becoming so large or heavy that they obscured the landscape or crowded one another. Each was positioned and adjusted individually, while still following the same system throughout the map.

After completing the Middle‑earth map, I turned to Beleriand. It was a particularly enjoyable map to work on because the geography is so different from anything in the real world, while also being more difficult to interpret. There is no single fixed map that answers every question, and the surviving sources do not always agree.

I wanted to show more than Beleriand alone, including the northern lands around Angband and the connection eastwards across Ered Luin towards Eriador. This required careful comparison of the available maps and some judgement in the areas where the wider geography is only partly shown.

Beleriand was developed at the same geographic scale as Middle‑earth, at 1:3,000,000. This made it possible to compare the two directly and later became important when I brought them together in one larger map.

The Shire map came from a different desire. It is where the journey begins with the hobbits, and I wanted to see more of that landscape at a closer scale. Together with Middle‑earth and Beleriand, it also felt like one of the three principal maps I most wanted to make.

The Middle‑earth DEM was too coarse for the Shire’s scale of 1:500,000, although I could use it as a soft base for the new and much more detailed Shire DEM. The challenge was not scale alone. The Shire has none of the great mountain ranges that naturally dominate a map, so the landscape had to remain interesting through much smaller changes in the land.

Although the maps were made digitally, I always intended them to become physical works. So much of our lives now takes place through screens, and for me there was something important in bringing the maps back into the physical world. Tolkien’s own work was created through writing, drawing and careful revision by hand. I cannot know what he would have thought of my maps, yet I felt that a printed object was more in keeping with that human quality than leaving the work only as a digital image. I explain more about the papers and printing in About the Prints.

The maps were designed to be read from across a room, while still holding enough detail to reward a much closer look. I chose to build them at twice the resolution required for the prints, a decision that proved useful when I later began preparing the maps for the book. The resulting files became too large to keep as one conventional stacked Photoshop document, so many of the elements had to be maintained separately in dedicated files. The working folder for the Tolkien maps now exceeds one terabyte, excluding the original elevation datasets.

The book came from a desire to offer a closer way of reading the maps. A wall print allows the geography to be seen as a whole, but many smaller areas cannot be studied at the same scale. In The Maps of Middle‑earth, the Beleriand, Middle‑earth and Shire maps could be divided into more detailed views and given the space needed for slower and closer inspection.

I do not release the full-resolution digital map files. A great deal of time, care and personal commitment has gone into developing them, and keeping control of the source material is also necessary to protect the work. Just as importantly, the maps were made for a particular physical result. By printing them myself, I can control the final quality in a way that would not be possible once the files were passed on to someone else. The digital files are essential to the process, but for me the finished maps are the prints and the book.

Once I had completed both Beleriand and Middle‑earth, I thought it might not be too much work to bring them together in one combined Beleriand and Middle‑earth map. It would allow the two regions to be seen in relation to one another and show more clearly how Beleriand connected with the lands to the east.

There is, however, very little information about the eastern parts of Middle‑earth during the First Age. I therefore chose to show Beleriand in the First Age together with Middle‑earth as it appears in the late Third Age. The map is not intended to present both regions at one moment in history, but to show their geographic relationship at the shared scale of 1:3,000,000.

The combined map required a DEM of its own. I could not simply place the two existing DEMs beside one another, as the transition between them had to work as one landscape and the scale, hillshading and visual treatment had to remain consistent across the whole.

While developing this new DEM in 2026, I also brought the individual Middle‑earth and Beleriand maps into closer alignment with it. Some aspects of the terrain and print preparation were refined so that all three maps belonged more consistently to the same series.

The geography and edition sizes did not change, and the Limited Edition numbering continued. These were refinements made during the development of the combined map, rather than new editions of the existing works.

A great deal of time, thought and love has gone into these maps. What began as a wish to see more of Middle‑earth gradually became a much larger body of work than I had first imagined.

For me, the most important thing was always to approach the work with respect. These are my interpretations of lands created by someone else, and my contribution to them is a modest one. I have tried to show more of the landscape without losing sight of the source from which it came.

What began with one Middle‑earth terrain map gradually grew to include Beleriand and the Shire, followed by the larger combined maps. These works were later brought together in The Maps of Middle‑earth, where the maps can be explored more closely and in greater detail than is possible from a wall print alone.

I am very glad that the maps have found their way into the homes of people who care for Tolkien’s work as deeply as I do. It is always a particular joy when someone sends me a photograph of a framed print on the wall, especially after choosing the place where it belongs. After spending so many years looking at these landscapes, it is rather wonderful to see where they have travelled.