3f:null 40:["$","title",null,{"children":"Chernobyl Fallout: Distance from the Reactor vs. Where It Rained | Dusty Wolfe Synek | Observable"}] 46:Tdf3,# Chernobyl Fallout: Distance from the Reactor vs. Where It Rained ### Introduction & Hypothesis On 26 April 1986 the Chernobyl reactor released cesium-137 (Cs-137) into the air, and over the next two weeks winds carried it across Europe. For two days the Soviet Union said nothing. Then on the morning of 28 April, workers arriving at the Forsmark nuclear plant in Sweden, about 1,200 km away, set off its radiation alarms. The plant found no leak of its own, and a Swedish meteorologist traced the winds backward to the southwestern Soviet Union. Sweden's embassy in Moscow asked twice whether there had been an accident and was told no both times. That evening the Soviet Union announced the accident ([Kyiv Post](https://www.kyivpost.com/post/74677)). Fallout reaches the ground in two ways: dry deposition, where particles settle out of the air, and wet deposition, where rain washes them out. The simplest expectation is that the ground gets less contaminated the farther it is from the reactor. This study visually tests whether distance explains where the Cs-137 landed, or whether rain does. **Hypothesis Statement:** In the simulated Chernobyl release, Cs-137 deposition across Europe does not fall off steadily with distance from the reactor. Beyond about 500 km, the heaviest deposition lands in separate patches where the plume met rain, so wet deposition makes up most of the total in those cells. **In plain terms:** Chernobyl's fallout did not spread like a ripple that fades evenly the farther you go. Far from the reactor, the worst-hit places were wherever it happened to rain while the radioactive cloud was overhead. ### Data Sources & Scope * **Deposition and air concentration:** a simulation by the Swedish Meteorological and Hydrological Institute (SMHI) using its MATCH model, published as netCDF ([DOI 10.5879/ft2b-vt32](https://researchdata.se/en/catalogue/dataset/ecds0062-1), CC BY 4.0; Langner et al. 1998). It is a raster of 79 × 79 cells, each about 50 km across, with one layer per hour from 25 April to 10 May 1986 (354 layers). This is a model, not a set of measurements. The 1986 readings come from scattered monitoring stations, not a grid, so a model is the only way to get a value in every cell; it fills them from recorded wind and rainfall. Its authors report that the simulated cloud's extent, arrival time and duration agree well with observations, and when I checked it against 1,038 air samples from 80 stations, the places and times the model rates high were also high in the measurements (rank correlation +0.69). The exact amounts are rougher, so every statement below is about the simulated plume. * **Geographic context:** country outlines from [Natural Earth via world-atlas](https://github.com/topojson/world-atlas), drawn in the model's own rotated-pole projection so the grid cells stay square. ### The Narrative: What the Visualizations Show A straight plot of the deposition raster shows almost nothing: one cell at the reactor holds 310 kBq/m² while the median cell holds under 1, so a linear color scale paints nearly all of Europe the same color. The final views use a logarithmic, color-blind-safe scale (reversed cividis, darker means more) on a map with distance rings, so all five orders of magnitude are visible at once. Splitting the total into wet and dry deposition on one shared scale answers the rain question directly, and a scatter plot shows each cell's deposit against its distance. Sliders and tooltips let the reader move through time, change what counts as "far", and read exact values.47:T6cf,view3 = { const points = cells.filter(d => d.total >= 0.01 && d.km <= 2750); return Plot.plot({ width: Math.min(width, 980), height: 420, marginLeft: 60, title: "Deposition at the same distance varies by a factor of hundreds", subtitle: `Shaded region: the ${hotStats.far.toLocaleString()} cells beyond ${farKm.toLocaleString()} km${farKm === 500 ? ", the distance in the hypothesis" : ""}.`, x: {label: "Distance from the reactor (km)", domain: [0, 2750]}, y: {type: "log", label: "Cs-137 deposited (kBq/m², log scale)", domain: [0.01, 400], ticks: [0.01, 0.1, 1, 10, 100], tickFormat: ","}, color: {legend: true, domain: ["Mostly wet (rain)", "Mostly dry"], range: [wetColor, dryColor]}, marks: [ Plot.rect([0], {x1: farKm, x2: 2750, y1: 0.01, y2: 400, fill: "#000", fillOpacity: 0.05}), // tooltip: values for the nearest dot Plot.dot(points, {x: "km", y: "total", r: 1.6, fillOpacity: 0.55, fill: d => d.wet > d.dry ? "Mostly wet (rain)" : "Mostly dry", tip: true, title: d => `${d3.format(".3~r")(d.total)} kBq/m²\n${Math.round(d.km).toLocaleString()} km from the reactor\n${d3.format(".0%")(d.wet / (d.wet + d.dry))} came down with rain`}), Plot.line(bandMedians, {x: "km", y: "median", stroke: "#000", strokeWidth: 2}), Plot.text(bandMedians.slice(0, 1), {x: "km", y: "median", text: ["Median of each 250 km band"], dx: 8, dy: -8, textAnchor: "start", fontWeight: "bold", stroke: "#fff", strokeWidth: 3, fill: "#000"}), Plot.ruleY([hotStats.threshold], {x1: farKm, x2: 2750, stroke: "#000", strokeDasharray: "3,3"}), Plot.text([[2750, hotStats.threshold]], {text: ["Hot spots: top 10% of these cells"], dy: -7, textAnchor: "end"}) ] }); }48:T9bd,### Was the Hypothesis Supported? Mostly. Distance and rain behave as predicted, but the "separate patches" part holds only for the heaviest deposits: - **Deposition does not fall off steadily.** View 3 shows that cells at the same distance differ by a factor of hundreds, and View 1 shows heavy deposits 500 to 1,400 km out with lighter ground in between. - **The distant hot spots came from rain.** In View 2 the far patches appear on the wet map and are almost absent from the dry map. Beyond 500 km, the top 10% of cells received 85% of their Cs-137 with rain. - **"Separate patches" is only partly right.** At moderate levels (about 7 kBq/m²) most hot cells form one continuous area connected to the reactor. Only the heaviest deposits (15 kBq/m² and above) are separate, such as the eastern Black Sea coast and northern Greece. - **A caution on cause.** The model computes wet deposition from the rainfall data it was given, so this shows how the simulated plume behaves. It is not independent proof of what happened on the ground. ## Interactive components Each interaction is built into a plot and has a comment in the code saying what it does. The time slider under View 1 redraws the map at any of 59 six-hour steps. The threshold slider in View 3 moves the "far" boundary, and the shaded region, the hot-spot line and the rain-share sentence all recompute. Tooltips on Views 1 to 4 read out the exact values under the pointer, so the charts can stay uncluttered. ## Color justification and accessibility Deposition (Views 1 and 2) uses cividis, reversed so that darker means more. Deposition is an ordered quantity, so it is encoded with lightness, not hue, and cividis was designed to look nearly identical with and without red-green color blindness. The scale is logarithmic because the values span five orders of magnitude. Colors are drawn in six steps per tenfold change, which keeps the page fast; tooltips give exact values. Wet versus dry in View 3 is two categories, so it gets two hues: blue and orange from the Okabe-Ito palette, which stay distinct under red-green color blindness because they differ in both hue and lightness. Air concentration in View 4 uses the same reversed cividis, so darker means more in every view. The images below are screenshots of the four final views taken with Firefox's protanopia (red color blindness) simulation, under Accessibility, Simulate. The maps look almost unchanged, and the blue and orange dots in View 3 stay distinct.49:T4bf,**What is wrong with Revision 1** - **The data is hidden.** The color scale is linear, and one cell at the reactor holds 310 kBq/m² while the median cell holds under 1. Almost every cell lands in the bottom color, and only about 70 of the 6,241 cells stand out from it. In Tufte's terms the data density is very low for a graphic this size. - **The colormap misleads.** The default "turbo" rainbow varies in hue, and hue has no natural order, so the reader cannot tell which color means more without the legend. Ordered data needs a scale that runs steadily from light to dark. A rainbow also creates false boundaries where the hue changes quickly, and it is not safe for red-green color blindness. - **The shape is distorted.** The grid is 79 by 79 square cells, but the default plot is much wider than it is tall, so every cell is stretched sideways. Distances east to west look longer than the same distance north to south. - **It is not self-documenting.** There is no title, no units on the legend, and no way to tell this is Europe. **Next:** fix the scale (logarithmic), the colormap (ordered and color-blind safe), the proportions, and the legend. ## Revision 2. A scale and a colormap that show the data4a:T733,**What changed:** the color scale is logarithmic, so each step in color is a tenfold change and all five orders of magnitude are visible. The colormap is cividis, which increases steadily in lightness and was built to stay readable with red-green color blindness. The plot is square, so the cells are square. Cells under 0.01 kBq/m² are left blank instead of being painted as if they held data. **What is wrong with Revision 2** - **The axes mean nothing to a reader.** "Grid column 40" is a storage detail. Position on the page, the clearest way to show where something is, is spent on array indices instead of on location. - **No reference frame.** Without coastlines or the reactor's position, a viewer cannot tell where anything is, so the hypothesis (which is about distance from the reactor) cannot even be read off the picture. Tufte asks a graphic to make the viewer think about the substance of the data; here the viewer is stuck working out where they are. - **The legend is in the wrong language.** The default format prints 0.1 as "100m" (milli), which no reader of a fallout map will parse. - **Non-data ink.** The white gaps between cells form a grid that carries no information and shimmers against the dark cells. Tufte's advice is to erase ink that does not show data. - **Dark means "little".** Cividis runs from dark blue (low) to yellow (high). On a white page the large low-value area becomes the heaviest thing in the picture, and the hot spots are the lightest. **Next:** put the grid on a map, mark the reactor, and reconsider which end of the colormap is dark. ## Revision 3. On a map The model grid uses a rotated pole, so on an ordinary latitude and longitude plot its cells would be skewed. I draw the map in the grid's own rotated projection instead, so every cell stays a true square and nothing is resampled.4b:T46f,**What changed:** the raster sits on a map of Europe in the model's own projection, with the reactor marked. **What is wrong with Revision 3** - **The context is louder than the data.** Black 1-pixel borders over a dark colormap compete with the cells. Tufte's layering principle ("1 + 1 = 3 or more") says reference marks should sit quietly behind the data; here the borders are the strongest marks on the page. - **The red marker is a color-blindness trap.** Red on dark blue and olive is the kind of pairing that collapses under deuteranopia. - **Distance is still not readable.** The hypothesis is about distance from the reactor, and there is nothing on the map to judge it by. The reader would have to estimate 500 km by eye. - **Dark still means "little".** Same problem as Revision 2. - **The borders are today's, not 1986's.** That needs to be said on the graphic or the borders need to be so quiet that they read only as orientation. **Next:** mute the geography, reverse the colormap so darker means more, add distance rings, and replace the red marker. ## Revision 4. Quiet context, distance rings, and the rain question4c:T5b8,**What changed:** land is a pale gray, coastlines are thin, borders are white hairlines, and the colormap is reversed so the heaviest deposits are the darkest marks. Dashed rings every 500 km make distance readable directly. A second map asks the causal question: how much of each cell's deposit came with rain? **What is wrong with Revision 4** - **The rain-share map shows noise as if it were signal.** A cell with a trace of Cs-137 that happened to arrive in rain is painted as strongly as a real hot spot. A ratio of two tiny numbers is drawn with the same visual weight as a ratio of two large ones, which misrepresents the underlying values. That fails Tufte's test of graphical integrity, which asks that the visual effect be proportional to the quantity it shows. - **Red-yellow-green is the wrong palette.** It is a diverging scheme used on data with no meaningful midpoint, and red against green is the pairing that deuteranopia removes. Under the simulation the two ends become hard to tell apart. - **Two maps, two scales, no link.** The reader has to carry a location from one map to the other by eye. - **Nothing is interactive.** Exact values cannot be read, and the 500 km threshold in the hypothesis cannot be questioned. **Next:** replace the ratio map with wet and dry maps on one shared scale, add tooltips, and add a view that plots deposition against distance directly. ## Revision 5. Shared scales, tooltips, and a direct test of distance4d:T959,**What changed:** wet and dry deposition became two maps on one shared color scale (now View 2), so a patch that is dark on one and pale on the other answers the rain question without a ratio. Every map cell has a tooltip. This new chart plots each cell's deposit against its distance, which tests the hypothesis directly. **What is wrong with Revision 5** - **Overplotting.** About 5,300 opaque dots pile up; in the dense region the reader sees a blob and cannot judge how many cells are there. The data density is high, which Tufte favors, but the opaque dots hide one another, so the extra data cannot be read. - **The y axis runs down to 0.001.** That is far below anything that matters, so the bottom of the plot is spent on trace values. - **Half the color scale is wasted.** Plot centered the diverging purple-orange scheme on zero, so its legend runs from -1 to 1 and the whole purple half is never used. A share between 0 and 1 has no negative side. - **A continuous color scale answers a yes-or-no question.** The reader only needs to know whether a cell is mostly wet or mostly dry. Two categories only need two clearly different colors. - **There is no trend to compare against.** "Does not fall off steadily" has no baseline. - **The threshold is invisible.** The 500 km in the hypothesis is my choice and the reader cannot see or change it. **Next:** thin the dots, clip the axis, use two categorical colors, add the median of each distance band, and let the reader move the threshold. ## Revision 6. The final views The views at the top of this notebook are Revision 6. View 3 has translucent dots, a y axis that starts at 0.01, two color-blind-safe hues, a median line as the baseline, and a slider for the distance threshold that recomputes the rain share in the sentence below it. View 1 gained a time slider, so the reader can watch the patches appear when the plume meets rain instead of taking the final map on trust. It also gained direct labels on the main distant deposits. View 4 is new: it uses the time axis of the raster to show when the plume reached each distance. Views 3 and 4 have titles that state their finding, and every view has units on its scales and a written "What this shows" paragraph, so the finding does not depend on hovering. Accessibility is shown with protanopia screenshots of all four views, under "Color justification and accessibility".4e:T954,// map used by the final views. `values` has one number per cell (same order as `cells`), // `what` is the label for it in the tooltip depositionMap = (values, {width: w = Math.min(width, 720), title, subtitle, legend = true, tip = true, labels = true, forsmark = false, what = "deposited"} = {}) => Plot.plot({ width: w, projection, title, subtitle, caption: legend ? "Source: SMHI MATCH simulation (Langner et al. 1998). Cells under 0.01 kBq/m² are not drawn. Borders are present-day." : undefined, color: {...depositionColor, legend}, marks: [ Plot.geo(land, {fill: "#ececec"}), cellLayer(values), Plot.geo(borders, {stroke: "#fff", strokeWidth: 0.4, strokeOpacity: 0.7}), Plot.geo(land, {stroke: "#555", strokeWidth: 0.5}), Plot.geo(ringKm.map(ring), {stroke: "#222", strokeWidth: 0.7, strokeDasharray: "3,3"}), Plot.text(ringLabels, {x: "lon", y: "lat", text: d => `${d3.format(",")(d.km)} km`, fontSize: 10, stroke: "#fff", strokeWidth: 3, fill: "#222"}), Plot.dot([reactor], {x: "lon", y: "lat", r: 4, fill: "#fff", stroke: "#000", strokeWidth: 1.5}), Plot.text([reactor], {x: "lon", y: "lat", text: "name", dy: -11, fontWeight: "bold", stroke: "#fff", strokeWidth: 3, fill: "#000"}), // the Swedish plant where the accident was first detected (see the intro) forsmark ? Plot.dot([{lon: 18.17, lat: 60.40}], {x: "lon", y: "lat", r: 2.5, fill: "#fff", stroke: "#000"}) : null, forsmark ? Plot.text([{lon: 18.17, lat: 60.40}], {x: "lon", y: "lat", text: ["Forsmark"], dx: 5, dy: 9, textAnchor: "start", fontSize: 10, stroke: "#fff", strokeWidth: 3, fill: "#000"}) : null, labels ? Plot.text(patchLabels, {x: "lon", y: "lat", text: "name", fontSize: 11, stroke: "#fff", strokeWidth: 3, fill: "#000"}) : null, // tooltip: the cell nearest the pointer tip ? Plot.tip(cells, Plot.pointer({ x: "lon", y: "lat", title: (d, i) => `${d3.format(".3~r")(values[i])} kBq/m² ${what}\n${Math.round(d.km).toLocaleString()} km from the reactor\nFinal total ${d3.format(".3~r")(d.total)} kBq/m², ${d.wet + d.dry > 0 ? d3.format(".0%")(d.wet / (d.wet + d.dry)) : "0%"} from rain\n${Math.abs(d.lat).toFixed(1)}°${d.lat >= 0 ? "N" : "S"}, ${Math.abs(d.lon).toFixed(1)}°${d.lon >= 0 ? "E" : "W"}` })) : null, Plot.frame({stroke: "#999"}) ] })4f:[{"id":"5273ef339ae3bcfbcc897644813df5a0ec389835ca0f4c56f991d3326d784c5542721ecd446c5c4e1473265621144c97dfb1b57279ede145186fddd8e5a23ff5","href":"https://static.observableusercontent.com/files/5273ef339ae3bcfbcc897644813df5a0ec389835ca0f4c56f991d3326d784c5542721ecd446c5c4e1473265621144c97dfb1b57279ede145186fddd8e5a23ff5","name":"chernobyl_extract.json","type":"application/json","lastModified":1791145021130,"size":2614186},{"id":"0fbea2509b96fa015367028bcf25cc19f8ab299c9f8e61e0bff3f40d48ef35064ff350b7eb88e44dbae7397fcfdbcecfc74476d78a97b9213c3d1666a49c5421","href":"https://static.observableusercontent.com/files/0fbea2509b96fa015367028bcf25cc19f8ab299c9f8e61e0bff3f40d48ef35064ff350b7eb88e44dbae7397fcfdbcecfc74476d78a97b9213c3d1666a49c5421","name":"view1-protanopia.png","type":"image/png","lastModified":1792126795845,"size":458166},{"id":"b77cd02996b1d86634074c18abe077253ee733b4faadbf384e3e6af892cba22cc8f79bd2f45f83e3deb7b4120e683a527cb9b950e0208e01a7d439f4f57ef49a","href":"https://static.observableusercontent.com/files/b77cd02996b1d86634074c18abe077253ee733b4faadbf384e3e6af892cba22cc8f79bd2f45f83e3deb7b4120e683a527cb9b950e0208e01a7d439f4f57ef49a","name":"view2-protanopia.png","type":"image/png","lastModified":1792126795901,"size":508912},{"id":"2f182416798294ca421740dbb225746af5a329380c63e06b10c4fed5388e393a23887ad4aec91cb78b1872f61e41cdb703c0536b4c101d5e3846856cc920fa33","href":"https://static.observableusercontent.com/files/2f182416798294ca421740dbb225746af5a329380c63e06b10c4fed5388e393a23887ad4aec91cb78b1872f61e41cdb703c0536b4c101d5e3846856cc920fa33","name":"view3-protanopia.png","type":"image/png","lastModified":1792126795955,"size":387388},{"id":"27dcd2b3692663cd5bf64180503be89eb6885cdf42e941238fa6cd387f44c66f262423258c039b5603636bb4b29a269a73cd3f768eda42281494c798c60293a1","href":"https://static.observableusercontent.com/files/27dcd2b3692663cd5bf64180503be89eb6885cdf42e941238fa6cd387f44c66f262423258c039b5603636bb4b29a269a73cd3f768eda42281494c798c60293a1","name":"view4-protanopia.png","type":"image/png","lastModified":1792126796010,"size":65008}] 44:{"id":"dbf380a12c23028a","slug":"chernobyl-fallout-distance-from-the-reactor-vs-where-it-rained","title":"Chernobyl Fallout: Distance from the Reactor vs. Where It Rained","thumbnail":"575b12761b1476b85f0463cbc6403dd57e1cfc70b26320e692099d4fb5ce7f6d","access_level":"public","create_time":"$D2026-10-15T18:20:17.822Z","update_time":"$D2026-10-16T05:08:48.461Z","view_time":null,"first_public_time":"$D2026-10-16T16:35:10.139Z","has_unpublished_edits":null,"fork":false,"stars":0,"starred":false,"owner":{"type":"team","id":"dff9b5e6e979adb0","login":"dlw272","name":"Dusty Wolfe Synek","avatar_url":null,"tier":"starter_2024","bio":""},"version":297,"publish_version":297,"role":null,"authored":false,"authors":[{"type":"individual","id":"318cf45c19986f2e","login":"dlw272","name":"Dusty Wolfe Synek","avatar_url":null,"tier":null,"bio":""}],"members":[{"type":"team","id":"dff9b5e6e979adb0","login":"dlw272","name":"Dusty Wolfe Synek","avatar_url":null,"tier":"starter_2024","role":"editor"}],"comments":[],"custom_thumbnail":false,"trashed":false,"lapsed":null,"subscribed":null,"forks":0,"collection":null,"type":"notebook","body":{"titling":"auto","title":"Chernobyl Fallout: Distance from the Reactor vs. Where It Rained","cells":[{"id":1,"value":"$46","pinned":false,"hidden":false,"mode":"md","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":2,"value":"---\n# Final views","pinned":false,"hidden":false,"mode":"md","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":4,"value":"## View 1. Where the Cs-137 landed","pinned":true,"hidden":false,"mode":"md","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":5,"value":"viewof frame = Inputs.range([0, data.time_6h.length - 1], {\n step: 1,\n value: data.time_6h.length - 1,\n label: \"Time step (6 hours each)\"\n})","pinned":true,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":6,"value":"// redraws when the time slider moves\nview1 = depositionMap(data.total_6h_kBq_m2[frame].flat(), {\n title: `Cs-137 on the ground by ${formatTime(data.time_6h[frame])}`,\n subtitle: \"Simulated total deposition. Dashed rings mark distance from the reactor.\",\n what: \"deposited by this time\",\n forsmark: true,\n labels: frame === data.time_6h.length - 1\n})","pinned":true,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":7,"value":"**What this shows:** the deposit does not spread outward as an even ring. Slide back to 28 April and only 3% of what the distant hot spots will receive has landed; by 3 May half of their final deposit has landed, and by 9 May 90% of it. Heavy patches appear 500 to 1,400 km out, in western Russia, Austria, central Sweden, southern Finland, northern Greece and on the eastern Black Sea coast, with lighter ground in between. In a typical distant hot spot about 40% of the final deposit arrived in a single six-hour step, consistent with a passing rain shower.","pinned":true,"hidden":false,"mode":"md","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":64,"value":"## View 2. Rain versus no rain\nThe same total, split by how it reached the ground. Both maps share one color scale, so they can be compared cell for cell.\n","pinned":true,"hidden":false,"mode":"md","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":8,"value":"view2 = html`

\n ${depositionMap(cells.map(d => d.wet), {width: Math.min(width, 980) / 2 - 6, title: \"Wet deposition (washed out by rain)\", what: \"came down with rain\", legend: false, labels: false})}\n ${depositionMap(cells.map(d => d.dry), {width: Math.min(width, 980) / 2 - 6, title: \"Dry deposition (settled without rain)\", what: \"settled dry\", legend: false, labels: false})}\n

${depositionLegend()}

\n ${await Promise.all(shots.map(async ([name, file]) => html`

`))}\n

`;\n}","pinned":true,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":17,"value":"---\n## Exploration and the Visual Editing Process\nSix revisions led to the views above. Each one is followed by what was wrong with it and what the next one changes.\n\n## Revision 1. Straight from the library example\nThe starting point is Observable Plot's raster example with my array dropped in: default colors, default scale, no geography.","pinned":true,"hidden":false,"mode":"md","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":18,"value":"revision1 = Plot.plot({\n color: {legend: true},\n marks: [Plot.raster(data.total_kBq_m2.flat(), {width: 79, height: 79})]\n})","pinned":true,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":19,"value":"$49","pinned":true,"hidden":false,"mode":"md","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":20,"value":"revision2 = Plot.plot({\n width: 560,\n height: 560,\n x: {label: \"Grid column (west to east)\", ticks: d3.range(0, 79, 10)},\n y: {label: \"Grid row (south to north)\", ticks: d3.range(0, 79, 10), reverse: true},\n color: {type: \"log\", scheme: \"cividis\", legend: true, label: \"Cs-137 deposited (kBq/m², log scale)\"},\n marks: [\n Plot.cell(cells.filter(d => d.total >= 0.01), {x: \"col\", y: \"row\", fill: \"total\"}),\n Plot.frame()\n ]\n})","pinned":true,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":21,"value":"$4a","pinned":true,"hidden":false,"mode":"md","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":22,"value":"revision3 = Plot.plot({\n width: 640,\n projection,\n color: {type: \"log\", scheme: \"cividis\", legend: true, label: \"Cs-137 deposited (kBq/m², log scale)\"},\n marks: [\n cellLayer(cells.map(d => d.total)),\n Plot.geo(countries, {stroke: \"black\", strokeWidth: 1}),\n Plot.dot([reactor], {x: \"lon\", y: \"lat\", fill: \"red\", r: 5}),\n Plot.text([reactor], {x: \"lon\", y: \"lat\", text: \"name\", dy: -12, fill: \"red\", fontSize: 14}),\n Plot.frame()\n ]\n})","pinned":true,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":23,"value":"$4b","pinned":true,"hidden":false,"mode":"md","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":24,"value":"revision4 = html`

\n ${depositionMap(cells.map(d => d.total), {width: 470, title: \"Total Cs-137 deposited\", tip: false, labels: false})}\n ${Plot.plot({\n width: 470,\n projection,\n title: \"Share that came down with rain\",\n color: {type: \"linear\", scheme: \"RdYlGn\", domain: [0, 1], legend: true, label: \"Wet share of deposition\", tickFormat: \"%\"},\n marks: [\n Plot.geo(land, {fill: \"#ececec\"}),\n cellLayer(cells.map(d => d.wet + d.dry > 0 ? d.wet / (d.wet + d.dry) : NaN), v => Math.round(v * 20) / 20),\n Plot.geo(land, {stroke: \"#555\", strokeWidth: 0.5}),\n Plot.frame({stroke: \"#999\"})\n ]\n })}\n

`","pinned":true,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":25,"value":"$4c","pinned":true,"hidden":false,"mode":"md","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":26,"value":"revision5 = Plot.plot({\n width: 640,\n height: 380,\n marginLeft: 60,\n x: {label: \"Distance from the reactor (km)\"},\n y: {type: \"log\", label: \"Cs-137 deposited (kBq/m²)\"},\n color: {legend: true, scheme: \"PuOr\", label: \"Wet share\"},\n marks: [\n Plot.dot(cells.filter(d => d.total > 0), {x: \"km\", y: \"total\", fill: d => d.wet / (d.wet + d.dry), r: 2, tip: true})\n ]\n})","pinned":true,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":27,"value":"$4d","pinned":true,"hidden":false,"mode":"md","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":28,"value":"---\n# Appendix: data and helper cells","pinned":true,"hidden":false,"mode":"md","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":29,"value":"data = FileAttachment(\"chernobyl_extract.json\").json()","pinned":true,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":30,"value":"// one object per grid cell (row 0 is the southern edge)\ncells = data.rlat.flatMap((rlat, row) => data.rlon.map((rlon, col) => ({\n col,\n row,\n lon: data.lon[row][col],\n lat: data.lat[row][col],\n km: data.distance_km[row][col],\n total: data.total_kBq_m2[row][col],\n wet: data.wet_kBq_m2[row][col],\n dry: data.dry_kBq_m2[row][col]\n})))","pinned":true,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":31,"value":"reactor = ({name: \"Chernobyl\", lon: 30.099, lat: 51.389})","pinned":true,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":32,"value":"// the model's grid is 0.5 degrees on a globe with the north pole moved to 30°N, 180°W.\n// an equirectangular projection rotated to 0°E, 60°N undoes that, so every cell is a square\nrotation = d3.geoRotation([0, -60])","pinned":true,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":33,"value":"// grid outline (cell edges), used to fit the map to the raster\ngridOutline = {\n const half = (data.rlon[1] - data.rlon[0]) / 2;\n const [x0, x1] = [data.rlon[0] - half, data.rlon[data.rlon.length - 1] + half];\n const [y0, y1] = [data.rlat[0] - half, data.rlat[data.rlat.length - 1] + half];\n const t = d3.range(0, 1.0001, 0.02);\n const edge = [\n ...t.map(k => [x0 + k * (x1 - x0), y0]),\n ...t.map(k => [x1, y0 + k * (y1 - y0)]),\n ...t.map(k => [x1 - k * (x1 - x0), y1]),\n ...t.map(k => [x0, y1 - k * (y1 - y0)])\n ];\n return {type: \"MultiPoint\", coordinates: edge.map(p => rotation.invert(p))};\n}","pinned":true,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":34,"value":"// one square per cell, same order as `cells`. corners are worked out in the rotated grid and\n// converted to lon/lat, so cells sit exactly where the model put them (no resampling)\ncellShapes = {\n const half = (data.rlon[1] - data.rlon[0]) / 2;\n return data.rlat.flatMap(y => data.rlon.map(x => ({\n type: \"Polygon\",\n coordinates: [[[x - half, y - half], [x - half, y + half], [x + half, y + half], [x + half, y - half], [x - half, y - half]]\n .map(p => rotation.invert(p))]\n })));\n}","pinned":true,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":35,"value":"projection = ({type: \"equirectangular\", rotate: [0, -60], domain: gridOutline})","pinned":true,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":36,"value":"topojson = import(\"https://cdn.jsdelivr.net/npm/topojson-client@3/+esm\")","pinned":true,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":37,"value":"// Natural Earth 1:50m from world-atlas. borders are today's, not 1986's\nworld = d3.json(\"https://cdn.jsdelivr.net/npm/world-atlas@2/countries-50m.json\")","pinned":true,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":38,"value":"land = topojson.merge(world, nearby)","pinned":true,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":39,"value":"countries = topojson.mesh(world, {type: \"GeometryCollection\", geometries: nearby})","pinned":true,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":40,"value":"borders = topojson.mesh(world, {type: \"GeometryCollection\", geometries: nearby}, (a, b) => a !== b)","pinned":true,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":41,"value":"// color: reversed cividis on a log scale, darker = more.\n// it reads the same with or without red-green color blindness\ndepositionColor = ({\n type: \"log\",\n scheme: \"cividis\",\n reverse: true,\n domain: [0.01, 310],\n clamp: true,\n label: \"Cs-137 deposited (kBq/m², log scale)\",\n ticks: [0.01, 0.1, 1, 10, 100],\n tickFormat: \",\"\n})","pinned":true,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":42,"value":"// same ramp for air so darker = more in every view\nairColor = ({\n type: \"log\",\n scheme: \"cividis\",\n reverse: true,\n domain: [0.001, 100],\n clamp: true,\n label: \"Cs-137 in air near the ground (Bq/m³, log scale)\",\n ticks: [0.001, 0.01, 0.1, 1, 10, 100],\n tickFormat: \",\"\n})","pinned":true,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":43,"value":"// wet vs dry: two Okabe-Ito colors that also differ in lightness\nwetColor = \"#0072B2\"","pinned":true,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":44,"value":"dryColor = \"#E69F00\"","pinned":true,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":45,"value":"// circle of radius km around the reactor\nring = km => d3.geoCircle().center([reactor.lon, reactor.lat]).radius(km / 6371 * 180 / Math.PI).precision(1)()","pinned":true,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":46,"value":"ringKm = [500, 1000, 1500, 2000]","pinned":true,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":47,"value":"// ring labels go on the south-west side of each ring\nringLabels = ringKm.map(km => {\n const [lon, lat] = d3.geoRotation([-reactor.lon, -reactor.lat]).invert(\n [-(km / 6371 * 180 / Math.PI) * Math.SQRT1_2, -(km / 6371 * 180 / Math.PI) * Math.SQRT1_2]\n );\n return {km, lon, lat};\n})","pinned":true,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":48,"value":"// labels for the big distant deposits (spots found in the patch analysis)\npatchLabels = [\n {name: \"Eastern Black\\nSea coast\", lon: 38.1, lat: 42.4},\n {name: \"Northern Greece\", lon: 23.1, lat: 40.3},\n {name: \"Western Russia\", lon: 37.4, lat: 53.6},\n {name: \"Austria\", lon: 14.3, lat: 47.9},\n {name: \"Central Sweden\", lon: 16.8, lat: 60.9},\n {name: \"Southern Finland\", lon: 23.0, lat: 61.4}\n]","pinned":true,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":49,"value":"formatTime = t => d3.utcFormat(\"%-d %B %H:%M UTC\")(new Date(t))","pinned":true,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":50,"value":"depositionLegend = () => Plot.legend({color: {...depositionColor, ticks: [0.01, 0.1, 1, 10, 100]}, width: 320})","pinned":true,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":51,"value":"$4e","pinned":true,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":52,"value":"// median per 250 km band, i.e. what a pure distance fall-off would look like\nbandMedians = d3.rollups(\n cells.filter(d => d.km <= 2750),\n v => d3.median(v, d => d.total),\n d => Math.floor(d.km / data.band_km)\n).map(([band, median]) => ({km: (band + 0.5) * data.band_km, median}))\n .filter(d => d.median >= 0.01)\n .sort((a, b) => a.km - b.km)","pinned":true,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":53,"value":"// reruns when the distance slider moves\nhotStats = {\n const far = cells.filter(d => d.km > farKm && d.total > 0);\n const threshold = d3.quantile(far.map(d => d.total).sort(d3.ascending), 0.9);\n const share = v => d3.sum(v, d => d.wet) / d3.sum(v, d => d.wet + d.dry);\n return {\n far: far.length,\n threshold,\n hotWet: share(far.filter(d => d.total >= threshold)),\n otherWet: share(far.filter(d => d.total < threshold))\n };\n}","pinned":true,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":54,"value":"// air concentration in the lowest model layer, by distance band and hour (View 4)\narrival = data.surface_conc_Bq_m3_by_band_hour.flatMap((row, band) =>\n band * data.band_km >= 2750 ? [] : row.map((conc, hour) => {\n const start = new Date(data.time[hour]);\n return {km: band * data.band_km, start, end: new Date(+start + 3600e3), conc: conc >= 0.001 ? conc : NaN};\n })\n)","pinned":true,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":59,"value":"// cells in the same color step get merged into one path, so a map is a few dozen paths\n// instead of ~6,000 and the page stays fast. `step` rounds a value to its color step\n// (6 steps per 10x by default). tooltips still show the exact value\ncellLayer = (values, step = v => v >= depositionColor.domain[0]\n ? 10 ** (Math.round(Math.log10(Math.min(v, depositionColor.domain[1])) * 6) / 6)\n : NaN) => {\n const groups = d3.group(d3.range(values.length).filter(i => !isNaN(step(values[i]))), i => step(values[i]));\n return Plot.geo(\n Array.from(groups, ([value, index]) => ({type: \"MultiPolygon\", coordinates: index.map(i => cellShapes[i].coordinates), value})),\n {fill: \"value\", stroke: \"value\", strokeWidth: 0.3}\n );\n}","pinned":true,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":60,"value":"// only countries within 5,000 km. drawing the whole world on every map was slow\nnearby = world.objects.countries.geometries.filter(g =>\n d3.geoDistance(d3.geoCentroid(topojson.feature(world, g)), [reactor.lon, reactor.lat]) * 6371 < 5000)","pinned":true,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"}],"files":"$W4f","stdlib":"2"}} 45:"$44:body"