31:T4fe,## Overview The netCDF Reader is a helper notebook for the users who want to customize the EMC Observable notebooks or develop their own Observable notebooks to work with the EMC-compatible netCDF files. This notebook currently supports netCDF 3 format and provides utility functions and guides for reading and extracting information from the netCDF files in Observable notebooks. This notebook uses the [netcdfjs](https://cheminfo.github.io/netcdfjs/) package to read netCDF files and provides the following functions to work with the model data:
- **`getMetadata`**(*reader*) returns model’s metadata. - **`modelURL`**(*reader*) returns model’s detail page URL. - **`getDimensions`**(*reader*) returns dimensions of variables. - **`getVariables`**(*reader*) returns all variables along with their metadata. - **`readRawData`**(*reader*) reads and returns data as stored in netcdf file. - **`extractVariableData`**(*reader*, *par*) read and returns data associated to a variable. - **`extractParameterData`**(*reader*, *key*, *param*, *value*) extract data for a model parameter at the given dimension variable. ## Input File On invocation, this notebook automatically loads a default Earth model file, **${defaultModel.name}**, but could also read a local file.32:T420,function getVariables(reader) { let data = { dimension_variables: {}, model_variables: {}, all_variables: {} }; let vars = []; const all_tag = "all_variables"; const headerVariables = reader.header.variables; const arrayLength = headerVariables.length; for (var i = 0; i < arrayLength; i++) { const key = headerVariables[i].name; vars.push(key); let varTag = "model_variables"; if (headerVariables[i].dimensions.length == 1) { varTag = "dimension_variables"; } data[varTag][key] = {}; data[all_tag][key] = {}; data[varTag][key]["dimensions"] = headerVariables[i].dimensions.map( (d) => vars[d] ); data[all_tag][key]["dimensions"] = data[varTag][key]["dimensions"]; const attrib = headerVariables[i].attributes; for (var j = 0; j < attrib.length; j++) { const akey = headerVariables[i].attributes[j]["name"]; data[varTag][key][akey] = headerVariables[i].attributes[j].value; data[all_tag][key][akey] = data[varTag][key][akey]; } } return data; }33:[{"id":"ce7557c4cf51004137ff8ad7fb4dfa53bec60401cc4b9892cebd5f37b7872ebab329eb2a82be5e458e2aee29a8e7b990c1d0c1a93a932b9d641b0b958c6d4b1c","href":"https://static.observableusercontent.com/files/ce7557c4cf51004137ff8ad7fb4dfa53bec60401cc4b9892cebd5f37b7872ebab329eb2a82be5e458e2aee29a8e7b990c1d0c1a93a932b9d641b0b958c6d4b1c","name":"BBNAP19-MOD-3D (1).nc","type":"application/octet-stream","lastModified":1678685640224.855,"size":20203240}] 2f:{"title":"EMC netCDF Reader","cells":[{"id":1028,"value":"# EMC netCDF Reader
\n**_Beta_**\n\n**Quick links**\n\n[→ EMC—Earth Model Collaboration](http://ds.iris.edu/ds/products/emc/)
\n[→ Repository Explorer—search and discover EMC models](https://observablehq.com/@earthscope/emc-repository-explorer?collection=@earthscope/models)
\n[→ Horizontal Slicer—quick visualization of a horizontal slice of an EMC model](https://observablehq.com/@earthscope/emc-horizontal-slicer)
","pinned":false,"hidden":false,"mode":"md"},{"id":1031,"value":"toc()","pinned":false,"hidden":false,"mode":"ojs"},{"id":0,"value":"$31","pinned":false,"hidden":false,"mode":"md"},{"id":21,"value":"The default input file is the **${defaultModel.name}** file from the EMC repository that is attached to this notebook.","pinned":false,"hidden":false,"mode":"md"},{"id":909,"value":"defaultModel = FileAttachment(\"BBNAP19-MOD-3D (1).nc\")","pinned":false,"hidden":false,"mode":"ojs"},{"id":38,"value":"💡 To override the default model file and read a local compatible netCDF file from your computer, click on the **Choose File** button below.\n","pinned":false,"hidden":false,"mode":"md"},{"id":14,"value":"viewof model = Inputs.file({\n label: \"Input Model\",\n accept: \"nc\",\n required: false\n})","pinned":false,"hidden":false,"mode":"ojs"},{"id":52,"value":"## Reading the Input File","pinned":false,"hidden":false,"mode":"md"},{"id":45,"value":"We use the *NetCDFReader* from the *netcdfjs* package to read the model netCDF file in the browser.","pinned":false,"hidden":false,"mode":"md"},{"id":43,"value":"netcdf = import(\"https://cdn.skypack.dev/netcdfjs@2.0.1?min\").then(\n (d) => d.NetCDFReader\n)","pinned":false,"hidden":false,"mode":"ojs"},{"id":59,"value":"reader = {\n let inputModel = model;\n if (inputModel === null) {\n inputModel = defaultModel;\n }\n\n return new netcdf(await inputModel.arrayBuffer());\n}","pinned":false,"hidden":false,"mode":"ojs"},{"id":97,"value":"## `getMetadata`","pinned":false,"hidden":false,"mode":"md"},{"id":407,"value":"Extract the metadata and return it as an object.","pinned":false,"hidden":false,"mode":"md"},{"id":102,"value":"function getMetadata(reader) {\n let metadata = {};\n const globalAttributes = reader.header.globalAttributes;\n const arrayLength = globalAttributes.length;\n for (var i = 0; i < arrayLength; i++) {\n const key = globalAttributes[i][\"name\"];\n metadata[key] = globalAttributes[i].value;\n }\n\n return metadata;\n}","pinned":false,"hidden":false,"mode":"ojs"},{"id":412,"value":"**Example**:","pinned":false,"hidden":false,"mode":"md"},{"id":69,"value":"metadata = getMetadata(reader)","pinned":false,"hidden":false,"mode":"ojs"},{"id":797,"value":"## `modelURL`","pinned":false,"hidden":false,"mode":"md"},{"id":800,"value":"Model detail usage based on the model DOI.","pinned":false,"hidden":false,"mode":"md"},{"id":790,"value":"function modelURL(reader) {\n // The default DOI is the EMC.\n let doi = \"10.17611/dp/emc.1\";\n let metadata = getMetadata(reader);\n if (\"repository_pid\" in metadata) {\n doi = metadata.repository_pid.split(\":\")[1];\n }\n return \"https://doi.org/\" + doi;\n}","pinned":false,"hidden":false,"mode":"ojs"},{"id":875,"value":"modelURL(reader)","pinned":false,"hidden":false,"mode":"ojs"},{"id":118,"value":"## `getDimensions`","pinned":false,"hidden":false,"mode":"md"},{"id":423,"value":"Get dimensions of the variables.","pinned":false,"hidden":false,"mode":"md"},{"id":107,"value":"// Get the variable dimensions.\nfunction getDimensions(reader) {\n const dims = reader.header.dimensions;\n let dict = {};\n dims.map((d) => (dict[d.name] = d.size));\n // dict[\"variables\"] = Object.keys(dict);\n return dict;\n}","pinned":false,"hidden":false,"mode":"ojs"},{"id":415,"value":"**Example**:","pinned":false,"hidden":false,"mode":"md"},{"id":120,"value":"dimensions = getDimensions(reader)","pinned":false,"hidden":false,"mode":"ojs"},{"id":153,"value":"## `getVariables`","pinned":false,"hidden":false,"mode":"md"},{"id":427,"value":"This function gets information on both dimension and model variables. The resulting Object contains all variables and their associated metadata.","pinned":false,"hidden":false,"mode":"md"},{"id":139,"value":"$32","pinned":false,"hidden":false,"mode":"ojs"},{"id":429,"value":"**Example**:","pinned":false,"hidden":false,"mode":"md"},{"id":144,"value":"variables = getVariables(reader)","pinned":false,"hidden":false,"mode":"ojs"},{"id":291,"value":"## `readRawData`","pinned":false,"hidden":false,"mode":"md"},{"id":437,"value":"Extract data as stored in the netCDF file.","pinned":false,"hidden":false,"mode":"md"},{"id":231,"value":"function readRawData(reader) {\n let data = {};\n const vars = getVariables(reader).all_variables;\n for (const key in vars) {\n data[key] = {};\n const dims = vars[key].dimensions;\n data[key][\"dimensions\"] = dims;\n data[key][\"values\"] = reader.getDataVariable(key);\n if (\"missing_value\" in Object.keys(vars[key])) {\n if (vars[key].missing_value != NaN) {\n for (let i = 0; i < data[key].length; i++) {\n if (data[key][i] == \"\" || data[key][i] == vars[key].missing_value) {\n data[key][i] = NaN;\n }\n }\n }\n }\n }\n return data;\n}","pinned":false,"hidden":false,"mode":"ojs"},{"id":440,"value":"**Example**:","pinned":false,"hidden":false,"mode":"md"},{"id":257,"value":"rawData = readRawData(reader)","pinned":false,"hidden":false,"mode":"ojs"},{"id":293,"value":"## `extractVariableData`","pinned":false,"hidden":false,"mode":"md"},{"id":451,"value":"Extract data for a model variable.","pinned":false,"hidden":false,"mode":"md"},{"id":295,"value":"function extractVariableData(reader, par) {\n const dims = getDimensions(reader);\n const vars = getVariables(reader).all_variables;\n const raw = readRawData(reader);\n let data = [];\n\n // a dimension variable, skip.\n if (par in Object.keys(dims)) {\n return data;\n }\n\n // Not a 3D variable, skip.\n const varDims = vars[par].dimensions;\n if (varDims.length != 3) {\n return data;\n }\n\n let l = 0;\n data[l] = {};\n\n // Loop over the first dimension.\n for (let i = 0; i < raw[varDims[0]].values.length; i++) {\n for (let j = 0; j < raw[varDims[1]].values.length; j++) {\n for (let k = 0; k < raw[varDims[2]].values.length; k++) {\n data[l][varDims[0]] = raw[varDims[0]].values[i];\n data[l][varDims[1]] = raw[varDims[1]].values[j];\n data[l][varDims[2]] = raw[varDims[2]].values[k];\n data[l][par] = raw[par].values[l];\n l++;\n data[l] = {};\n }\n }\n }\n return data;\n}","pinned":false,"hidden":false,"mode":"ojs"},{"id":453,"value":"**Example**:
\n**The first model variable.**","pinned":false,"hidden":false,"mode":"md"},{"id":563,"value":"var0 = Object.keys(getVariables(reader).model_variables)[0]","pinned":false,"hidden":false,"mode":"ojs"},{"id":390,"value":"var0Values = {\n return extractVariableData(reader, var0);\n}","pinned":false,"hidden":false,"mode":"ojs"},{"id":467,"value":"**Example**:","pinned":false,"hidden":false,"mode":"md"},{"id":460,"value":"## `extractParameterData`","pinned":false,"hidden":false,"mode":"md"},{"id":636,"value":"Extract data for a model parameter at the given dimension variable.","pinned":false,"hidden":false,"mode":"md"},{"id":392,"value":"function extractParameterData(reader, key, param, value) {\n const varData = extractVariableData(reader, key);\n const depthData = varData.filter((d) => +d[param] === value);\n return depthData;\n}","pinned":false,"hidden":false,"mode":"ojs"},{"id":520,"value":"**${var0} data for the tenth value of the first dimension.**","pinned":false,"hidden":false,"mode":"md"},{"id":580,"value":"dim0 = Object.keys(getVariables(reader).dimension_variables)[0]","pinned":false,"hidden":false,"mode":"ojs"},{"id":574,"value":"dim0Value = rawData[dim0].values[9]","pinned":false,"hidden":false,"mode":"ojs"},{"id":402,"value":"dim0Data = extractParameterData(reader, var0, dim0, dim0Value)","pinned":false,"hidden":false,"mode":"ojs"},{"id":524,"value":"**${var0} data for the sixth value of the second dimension.**","pinned":false,"hidden":false,"mode":"md"},{"id":618,"value":"dim1 = Object.keys(getVariables(reader).dimension_variables)[1]","pinned":false,"hidden":false,"mode":"ojs"},{"id":622,"value":"dim1Value = rawData[dim1].values[5]","pinned":false,"hidden":false,"mode":"ojs"},{"id":517,"value":"dim1Data = extractParameterData(reader, var0, dim1, dim1Value)","pinned":false,"hidden":false,"mode":"ojs"},{"id":529,"value":"**${var0} data for the eighteenth value of the second dimension.**","pinned":false,"hidden":false,"mode":"md"},{"id":533,"value":"dim2 = Object.keys(getVariables(reader).dimension_variables)[2]","pinned":false,"hidden":false,"mode":"ojs"},{"id":652,"value":"dim2Value = rawData[dim2].values[17]","pinned":false,"hidden":false,"mode":"ojs"},{"id":658,"value":"dim2Data = extractParameterData(reader, var0, dim2, dim2Value)","pinned":false,"hidden":false,"mode":"ojs"},{"id":887,"value":"## Plotting","pinned":false,"hidden":false,"mode":"md"},{"id":1017,"value":"The only *gotcha* that you really need to consider in taking a netCDF file and attempting to extract and plot data is sorting out how variables and dimensions are assigned.\n\nA 3D matrix representing netCDF data can be index by `depth`, `latitude`, and `longitude`... or `latitude`, `longitude`, and then `depth`... and so on.","pinned":false,"hidden":false,"mode":"md"},{"id":983,"value":"depths = rawData[dim2].values","pinned":false,"hidden":false,"mode":"ojs"},{"id":943,"value":"plotData = extractParameterData(reader, var0, dim2, depths[depth])","pinned":false,"hidden":false,"mode":"ojs"},{"id":1020,"value":"*This slider allows us to traverse this particular model's depths and render a variable color coded for each longitude and latitude at that depth.*\n\n***We have a horizontal \"depth slicer\" to visualize a netCDF model.***","pinned":false,"hidden":false,"mode":"md"},{"id":915,"value":"viewof depth = Inputs.range([0, depths.length - 1], {\n label: dim2,\n format: (d) => +depths[d],\n value: 0,\n step: 1\n})\n","pinned":false,"hidden":false,"mode":"ojs"},{"id":988,"value":"Plot.plot({\n projection: \"albers\",\n width,\n color: { scheme: \"turbo\", legend: true, nice: true, label: var0 },\n marks: [\n Plot.geo(nation),\n Plot.geo(states, { strokeOpacity: 0.25 }),\n Plot.raster(plotData, {\n clip: true,\n x: (d) => d.longitude,\n y: (d) => d.latitude,\n fill: (d) => d[var0],\n opacity: 0.5,\n interpolate: \"barycentric\",\n blur: 5\n })\n ]\n})","pinned":true,"hidden":false,"mode":"ojs"},{"id":1041,"value":"*Note the use of `Plot.raster` and barycentric interpolation to produce a pretty heatmap from regular lat/long grid samples.*","pinned":false,"hidden":false,"mode":"md"},{"id":892,"value":"---\n ## Appendix","pinned":false,"hidden":false,"mode":"md"},{"id":904,"value":"states = topojson.feature(us, us.objects.states)","pinned":false,"hidden":false,"mode":"ojs"},{"id":902,"value":"nation = topojson.feature(us, us.objects.nation)","pinned":false,"hidden":false,"mode":"ojs"},{"id":896,"value":"import { us } from \"@observablehq/us-geographic-data\"","pinned":false,"hidden":false,"mode":"ojs"},{"id":1034,"value":"import { toc } from \"@mbostock/toc\"","pinned":false,"hidden":false,"mode":"ojs"}],"files":"$W33","stdlib":"1","titling":"auto"}