From 3e01ae5ce72c02aa656e23769eac7b29049d1cd4 Mon Sep 17 00:00:00 2001 From: Tom White Date: Wed, 15 May 2024 12:33:36 +0100 Subject: [PATCH] Add notebooks for Pangeo examples at https://github.com/pangeo-data/distributed-array-examples --- .gitignore | 1 + examples/cubed.svg | 56 - examples/pangeo-1-vorticity.ipynb | 4140 ++++++++ examples/pangeo-2-quadratic-means.ipynb | 2463 +++++ examples/pangeo-3-tem.ipynb | 8318 +++++++++++++++++ .../pangeo-4-climatological-anomalies.ipynb | 2493 +++++ examples/pangeo-tem.ipynb | 2711 ------ examples/pangeo-vorticity.ipynb | 1318 --- examples/pangeo-vorticity.svg | 472 - 9 files changed, 17415 insertions(+), 4557 deletions(-) delete mode 100644 examples/cubed.svg create mode 100644 examples/pangeo-1-vorticity.ipynb create mode 100644 examples/pangeo-2-quadratic-means.ipynb create mode 100644 examples/pangeo-3-tem.ipynb create mode 100644 examples/pangeo-4-climatological-anomalies.ipynb delete mode 100644 examples/pangeo-tem.ipynb delete mode 100644 examples/pangeo-vorticity.ipynb delete mode 100644 examples/pangeo-vorticity.svg diff --git a/.gitignore b/.gitignore index edbbbd10..6e99b944 100644 --- a/.gitignore +++ b/.gitignore @@ -136,3 +136,4 @@ dmypy.json # Cubed .lithops_config +examples/cubed.svg diff --git a/examples/cubed.svg b/examples/cubed.svg deleted file mode 100644 index 313f6a51..00000000 --- a/examples/cubed.svg +++ /dev/null @@ -1,56 +0,0 @@ - - - - - - - -num tasks: 4 -max projected memory: 100.0 MB - - -array-001 - - -array-001 -asarray - - - - - -array-004 - - -array-004 -add (bw) - - - - - -array-001->array-004 - - - - - -array-002 - - -array-002 -asarray - - - - - -array-002->array-004 - - - - - diff --git a/examples/pangeo-1-vorticity.ipynb b/examples/pangeo-1-vorticity.ipynb new file mode 100644 index 00000000..72bf2aa7 --- /dev/null +++ b/examples/pangeo-1-vorticity.ipynb @@ -0,0 +1,4140 @@ +{ + "cells": [ + { + "cell_type": "markdown", + "id": "bdf4365e", + "metadata": {}, + "source": [ + "# Pangeo Vorticity Workload\n", + "\n", + "This is a notebook for exploring a simplified version of the example in https://github.com/pangeo-data/distributed-array-examples/issues/1.\n", + "\n", + "See also Tom Nicholas's [notebook](https://gist.github.com/TomNicholas/8366c917349b647d87860a20a257a3fb#file-benchmark-vorticity-ipynb) exploring this problem. Note that the code below for `diff` is based on the Dask version." + ] + }, + { + "cell_type": "code", + "execution_count": 1, + "id": "b8386400", + "metadata": {}, + "outputs": [], + "source": [ + "import cubed\n", + "import cubed.random\n", + "from cubed.extensions.rich import RichProgressBar\n", + "import xarray as xr" + ] + }, + { + "cell_type": "markdown", + "id": "9c1953a4", + "metadata": {}, + "source": [ + "Initialization parameters for the workload" + ] + }, + { + "cell_type": "code", + "execution_count": 2, + "id": "f4b61ed9", + "metadata": {}, + "outputs": [], + "source": [ + "t_length = 50\n", + "spec = cubed.Spec(allowed_mem=\"2GB\")" + ] + }, + { + "cell_type": "markdown", + "id": "48d462ed", + "metadata": {}, + "source": [ + "Create random data stored in Zarr." + ] + }, + { + "cell_type": "code", + "execution_count": 3, + "id": "11782285", + "metadata": {}, + "outputs": [ + { + "data": { + "application/vnd.jupyter.widget-view+json": { + "model_id": "40a37f546c3d4752a92531e87aeb4692", + "version_major": 2, + "version_minor": 0 + }, + "text/plain": [ + "Output()" + ] + }, + "metadata": {}, + "output_type": "display_data" + }, + { + "data": { + "text/html": [ + "
\n"
+      ],
+      "text/plain": []
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    },
+    {
+     "data": {
+      "text/html": [
+       "
\n",
+       "
\n"
+      ],
+      "text/plain": [
+       "\n"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    }
+   ],
+   "source": [
+    "u = cubed.random.random((t_length, 1, 987, 1920), chunks=(10, 1, -1, -1), spec=spec)\n",
+    "v = cubed.random.random((t_length, 1, 987, 1920), chunks=(10, 1, -1, -1), spec=spec)\n",
+    "dx = cubed.random.random((1, 987, 1920), chunks=(1, -1, -1), spec=spec)\n",
+    "dy = cubed.random.random((1, 987, 1920), chunks=(1, -1, -1), spec=spec)\n",
+    "arrays = [u, v, dx, dy]\n",
+    "paths = [\n",
+    "    f\"{spec.work_dir}/u_{t_length}.zarr\",\n",
+    "    f\"{spec.work_dir}/v_{t_length}.zarr\",\n",
+    "    f\"{spec.work_dir}/dx_{t_length}.zarr\",\n",
+    "    f\"{spec.work_dir}/dy_{t_length}.zarr\",\n",
+    "]\n",
+    "cubed.store(arrays, paths, compute_arrays_in_parallel=True, callbacks=[RichProgressBar()])"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 4,
+   "id": "58bbcfde",
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/html": [
+       "
\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "
<xarray.Dataset> Size: 2GB\n",
+       "Dimensions:  (time: 50, face: 1, j: 987, i: 1920)\n",
+       "Dimensions without coordinates: time, face, j, i\n",
+       "Data variables:\n",
+       "    U        (time, face, j, i) float64 758MB cubed.Array<chunksize=(10, 1, 987, 1920)>\n",
+       "    V        (time, face, j, i) float64 758MB cubed.Array<chunksize=(10, 1, 987, 1920)>\n",
+       "    dx       (face, j, i) float64 15MB cubed.Array<chunksize=(1, 987, 1920)>\n",
+       "    dy       (face, j, i) float64 15MB cubed.Array<chunksize=(1, 987, 1920)>
"
+      ],
+      "text/plain": [
+       " Size: 2GB\n",
+       "Dimensions:  (time: 50, face: 1, j: 987, i: 1920)\n",
+       "Dimensions without coordinates: time, face, j, i\n",
+       "Data variables:\n",
+       "    U        (time, face, j, i) float64 758MB cubed.Array\n",
+       "    V        (time, face, j, i) float64 758MB cubed.Array\n",
+       "    dx       (face, j, i) float64 15MB cubed.Array\n",
+       "    dy       (face, j, i) float64 15MB cubed.Array"
+      ]
+     },
+     "execution_count": 4,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "u = cubed.from_zarr(paths[0], spec=spec)\n",
+    "v = cubed.from_zarr(paths[1], spec=spec)\n",
+    "dx = cubed.from_zarr(paths[2], spec=spec)\n",
+    "dy = cubed.from_zarr(paths[3], spec=spec)\n",
+    "ds = xr.Dataset(\n",
+    "    dict(\n",
+    "        U=([\"time\", \"face\", \"j\", \"i\"], u),\n",
+    "        V=([\"time\", \"face\", \"j\", \"i\"], v),\n",
+    "        dx=([\"face\", \"j\", \"i\"], dx),\n",
+    "        dy=([\"face\", \"j\", \"i\"], dy),\n",
+    "    )\n",
+    ")\n",
+    "ds"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 5,
+   "id": "df8b97e8",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "def diff(da, dim):\n",
+    "    \"\"\"\n",
+    "    First order derivative along dim.\n",
+    "    \n",
+    "    Differencing removes one grid element, so pad to restore size.\n",
+    "    \"\"\"\n",
+    "    # Once https://github.com/cubed-dev/cubed/issues/475 is done we can use this simpler implmentation:\n",
+    "    # return da.diff(dim).pad({dim: (1, 0)}, mode='symmetric')\n",
+    "    \n",
+    "    ind = da.dims.index(dim)\n",
+    "    d = da.diff(dim).data\n",
+    "    pad_width = tuple((0, 0) if i != ind else (1, 0) for i in range(da.ndim))\n",
+    "    chunks = da.data.chunks # make sure result after padding has original chunk size\n",
+    "    p = cubed.pad(d, pad_width=pad_width, mode=\"symmetric\", chunks=chunks)\n",
+    "    return xr.DataArray(p, dims=da.dims)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 6,
+   "id": "c382a626",
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/html": [
+       "
\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "
<xarray.DataArray (face: 1, j: 987, i: 1920)> Size: 15MB\n",
+       "cubed.Array<array-061, shape=(1, 987, 1920), dtype=float64, chunks=((1,), (987,), (1920,))>\n",
+       "Dimensions without coordinates: face, j, i
"
+      ],
+      "text/plain": [
+       " Size: 15MB\n",
+       "cubed.Array\n",
+       "Dimensions without coordinates: face, j, i"
+      ]
+     },
+     "execution_count": 6,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "omega = diff(ds.U, dim='j') * ds.dx - diff(ds.V, dim='i') * ds.dy\n",
+    "mean = omega.mean('time', skipna=False)\n",
+    "mean"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 7,
+   "id": "403ae0b0",
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "image/svg+xml": [
+       "\n",
+       "\n",
+       "\n",
+       "num tasks: 51\n",
+       "max projected memory: 909.6 MB\n",
+       "total nbytes written: 3.9 GB\n",
+       "optimized: True\n",
+       "\n",
+       "\n",
+       "op-017\n",
+       "\n",
+       "\n",
+       "op-017\n",
+       "from_zarr\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "array-017\n",
+       "\n",
+       "\n",
+       "array-017\n",
+       "u\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "op-017->array-017\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "op-027\n",
+       "\n",
+       "\n",
+       "op-027\n",
+       "__sub__\n",
+       "tasks: 5\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "array-017->op-027\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "array-017->op-027\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "array-027\n",
+       "\n",
+       "\n",
+       "array-027\n",
+       "d\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "op-027->array-027\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "op-030\n",
+       "\n",
+       "\n",
+       "op-030\n",
+       "pad\n",
+       "tasks: 5\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "array-027->op-030\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "op-033\n",
+       "\n",
+       "\n",
+       "op-033\n",
+       "pad\n",
+       "tasks: 5\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "array-027->op-033\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "array-030\n",
+       "\n",
+       "\n",
+       "array-030\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "op-030->array-030\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "array-030->op-033\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "array-033\n",
+       "\n",
+       "\n",
+       "array-033\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "op-033->array-033\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "op-057\n",
+       "\n",
+       "\n",
+       "op-057\n",
+       "__sub__\n",
+       "tasks: 5\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "array-033->op-057\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "op-019\n",
+       "\n",
+       "\n",
+       "op-019\n",
+       "from_zarr\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "array-019\n",
+       "\n",
+       "\n",
+       "array-019\n",
+       "dx\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "op-019->array-019\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "op-037\n",
+       "\n",
+       "\n",
+       "op-037\n",
+       "__getitem__\n",
+       "tasks: 1\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "array-019->op-037\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "array-037\n",
+       "\n",
+       "\n",
+       "array-037\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "op-037->array-037\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "array-037->op-057\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "op-018\n",
+       "\n",
+       "\n",
+       "op-018\n",
+       "from_zarr\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "array-018\n",
+       "\n",
+       "\n",
+       "array-018\n",
+       "v\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "op-018->array-018\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "op-045\n",
+       "\n",
+       "\n",
+       "op-045\n",
+       "__sub__\n",
+       "tasks: 5\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "array-018->op-045\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "array-018->op-045\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "array-045\n",
+       "\n",
+       "\n",
+       "array-045\n",
+       "d\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "op-045->array-045\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "op-048\n",
+       "\n",
+       "\n",
+       "op-048\n",
+       "pad\n",
+       "tasks: 5\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "array-045->op-048\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "op-051\n",
+       "\n",
+       "\n",
+       "op-051\n",
+       "pad\n",
+       "tasks: 5\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "array-045->op-051\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "array-048\n",
+       "\n",
+       "\n",
+       "array-048\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "op-048->array-048\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "array-048->op-051\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "array-051\n",
+       "\n",
+       "\n",
+       "array-051\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "op-051->array-051\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "array-051->op-057\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "op-020\n",
+       "\n",
+       "\n",
+       "op-020\n",
+       "from_zarr\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "array-020\n",
+       "\n",
+       "\n",
+       "array-020\n",
+       "dy\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "op-020->array-020\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "op-055\n",
+       "\n",
+       "\n",
+       "op-055\n",
+       "__getitem__\n",
+       "tasks: 1\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "array-020->op-055\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "array-055\n",
+       "\n",
+       "\n",
+       "array-055\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "op-055->array-055\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "array-055->op-057\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "array-057\n",
+       "\n",
+       "\n",
+       "array-057\n",
+       "None\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "op-057->array-057\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "op-058\n",
+       "\n",
+       "\n",
+       "op-058\n",
+       "mean\n",
+       "tasks: 2\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "array-057->op-058\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "array-058\n",
+       "\n",
+       "\n",
+       "array-058\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "op-058->array-058\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "op-061\n",
+       "\n",
+       "\n",
+       "op-061\n",
+       "mean\n",
+       "tasks: 1\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "array-058->op-061\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "array-061\n",
+       "\n",
+       "\n",
+       "array-061\n",
+       "None\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "op-061->array-061\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "create-arrays\n",
+       "\n",
+       "\n",
+       "create-arrays\n",
+       "tasks: 11\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "arrays\n",
+       "\n",
+       "\n",
+       "arrays\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "create-arrays->arrays\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       ""
+      ],
+      "text/plain": [
+       ""
+      ]
+     },
+     "execution_count": 7,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "mean.data.visualize()"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 8,
+   "id": "7b8b6292",
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "application/vnd.jupyter.widget-view+json": {
+       "model_id": "ac32d115b19f4d9eb6e090a7cc9b6137",
+       "version_major": 2,
+       "version_minor": 0
+      },
+      "text/plain": [
+       "Output()"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    },
+    {
+     "data": {
+      "text/html": [
+       "
\n"
+      ],
+      "text/plain": []
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    },
+    {
+     "data": {
+      "text/html": [
+       "
\n",
+       "
\n"
+      ],
+      "text/plain": [
+       "\n"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    },
+    {
+     "data": {
+      "text/html": [
+       "
\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "
<xarray.DataArray (face: 1, j: 987, i: 1920)> Size: 15MB\n",
+       "array([[[-0.00935039,  0.05180967,  0.02368589, ...,  0.09296773,\n",
+       "         -0.02251309,  0.00811106],\n",
+       "        [-0.03833913,  0.05900023,  0.01459681, ..., -0.01437545,\n",
+       "          0.01901828,  0.02181335],\n",
+       "        [ 0.00342973,  0.01948303,  0.01160908, ...,  0.02988825,\n",
+       "         -0.0061075 , -0.01061161],\n",
+       "        ...,\n",
+       "        [ 0.03900664,  0.04362473,  0.00623131, ..., -0.03877337,\n",
+       "          0.00551559, -0.0664945 ],\n",
+       "        [-0.01084767, -0.00746532,  0.04994078, ...,  0.00983424,\n",
+       "         -0.00747178, -0.01885188],\n",
+       "        [ 0.00544092,  0.05520928, -0.0255991 , ...,  0.02842916,\n",
+       "          0.0289831 ,  0.01773723]]])\n",
+       "Dimensions without coordinates: face, j, i
"
+      ],
+      "text/plain": [
+       " Size: 15MB\n",
+       "array([[[-0.00935039,  0.05180967,  0.02368589, ...,  0.09296773,\n",
+       "         -0.02251309,  0.00811106],\n",
+       "        [-0.03833913,  0.05900023,  0.01459681, ..., -0.01437545,\n",
+       "          0.01901828,  0.02181335],\n",
+       "        [ 0.00342973,  0.01948303,  0.01160908, ...,  0.02988825,\n",
+       "         -0.0061075 , -0.01061161],\n",
+       "        ...,\n",
+       "        [ 0.03900664,  0.04362473,  0.00623131, ..., -0.03877337,\n",
+       "          0.00551559, -0.0664945 ],\n",
+       "        [-0.01084767, -0.00746532,  0.04994078, ...,  0.00983424,\n",
+       "         -0.00747178, -0.01885188],\n",
+       "        [ 0.00544092,  0.05520928, -0.0255991 , ...,  0.02842916,\n",
+       "          0.0289831 ,  0.01773723]]])\n",
+       "Dimensions without coordinates: face, j, i"
+      ]
+     },
+     "execution_count": 8,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "mean.compute(callbacks=[RichProgressBar()])"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "b9382d3b",
+   "metadata": {},
+   "source": [
+    "## A more efficient implementation\n",
+    "\n",
+    "In the above visualization we can see that the `diff` and `pad` operations are not fused together. This is something that could be improved in the future; see [#464](https://github.com/cubed-dev/cubed/issues/464) for what is needed.\n",
+    "\n",
+    "In the meantime, we can use Cubed's `map_overlap` to write a more efficient combined version of `diff` and `pad`, called `diff_pad`."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 9,
+   "id": "34f02045",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "from cubed.backend_array_api import namespace as nxp\n",
+    "from cubed.overlap import map_overlap\n",
+    "from cubed.vendor.dask.array.utils import validate_axis"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 10,
+   "id": "e839c886",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "def diff_pad(x, /, *, axis=-1):\n",
+    "    axis = validate_axis(axis, x.ndim)\n",
+    "    depth = tuple(1 if i == axis else 0 for i in range(x.ndim))\n",
+    "    return map_overlap(\n",
+    "        _diff_pad,\n",
+    "        x,\n",
+    "        dtype=x.dtype,\n",
+    "        chunks=x.chunks,\n",
+    "        depth=depth,\n",
+    "        boundary=0,  # actual value is ignored\n",
+    "        trim=False,\n",
+    "        axis=axis,\n",
+    "    )\n",
+    "\n",
+    "\n",
+    "def _diff_pad(a, block_id=None, axis=None):\n",
+    "    if block_id[axis] == 0:  # pad first block along axis\n",
+    "        sl = tuple(slice(1, -1) if i == axis else slice(None) for i in range(a.ndim))\n",
+    "        val = nxp.diff(a[sl], axis=axis)\n",
+    "        pad_width = tuple((1, 0) if i == axis else (0, 0) for i in range(a.ndim))\n",
+    "        val = nxp.pad(val, pad_width=pad_width, mode=\"symmetric\")\n",
+    "        return val\n",
+    "    else:\n",
+    "        sl = tuple(slice(None, -1) if i == axis else slice(None) for i in range(a.ndim))\n",
+    "        return nxp.diff(a[sl], axis=axis)"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "09bcf4eb",
+   "metadata": {},
+   "source": [
+    "Now we can write a more efficient version of `diff` from above"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 11,
+   "id": "c3fbcb1f",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "def diff2(da, dim):\n",
+    "    ind = da.dims.index(dim)\n",
+    "    dp = diff_pad(da.data, axis=ind)\n",
+    "    return xr.DataArray(dp, dims=da.dims)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 12,
+   "id": "45f3ebc1",
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/html": [
+       "
\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "
<xarray.DataArray (face: 1, j: 987, i: 1920)> Size: 15MB\n",
+       "cubed.Array<array-082, shape=(1, 987, 1920), dtype=float64, chunks=((1,), (987,), (1920,))>\n",
+       "Dimensions without coordinates: face, j, i
"
+      ],
+      "text/plain": [
+       " Size: 15MB\n",
+       "cubed.Array\n",
+       "Dimensions without coordinates: face, j, i"
+      ]
+     },
+     "execution_count": 12,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "omega = diff2(ds.U, dim='j') * ds.dx - diff2(ds.V, dim='i') * ds.dy\n",
+    "mean = omega.mean('time', skipna=False)\n",
+    "mean"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 13,
+   "id": "236caaa8",
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "image/svg+xml": [
+       "\n",
+       "\n",
+       "\n",
+       "num tasks: 27\n",
+       "max projected memory: 909.6 MB\n",
+       "total nbytes written: 2.4 GB\n",
+       "optimized: True\n",
+       "\n",
+       "\n",
+       "op-017\n",
+       "\n",
+       "\n",
+       "op-017\n",
+       "from_zarr\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "array-017\n",
+       "\n",
+       "\n",
+       "array-017\n",
+       "u\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "op-017->array-017\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "op-064\n",
+       "\n",
+       "\n",
+       "op-064\n",
+       "map_overlap\n",
+       "tasks: 5\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "array-017->op-064\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "array-064\n",
+       "\n",
+       "\n",
+       "array-064\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "op-064->array-064\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "op-078\n",
+       "\n",
+       "\n",
+       "op-078\n",
+       "__sub__\n",
+       "tasks: 5\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "array-064->op-078\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "op-019\n",
+       "\n",
+       "\n",
+       "op-019\n",
+       "from_zarr\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "array-019\n",
+       "\n",
+       "\n",
+       "array-019\n",
+       "dx\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "op-019->array-019\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "op-068\n",
+       "\n",
+       "\n",
+       "op-068\n",
+       "__getitem__\n",
+       "tasks: 1\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "array-019->op-068\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "array-068\n",
+       "\n",
+       "\n",
+       "array-068\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "op-068->array-068\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "array-068->op-078\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "op-018\n",
+       "\n",
+       "\n",
+       "op-018\n",
+       "from_zarr\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "array-018\n",
+       "\n",
+       "\n",
+       "array-018\n",
+       "v\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "op-018->array-018\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "op-072\n",
+       "\n",
+       "\n",
+       "op-072\n",
+       "map_overlap\n",
+       "tasks: 5\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "array-018->op-072\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "array-072\n",
+       "\n",
+       "\n",
+       "array-072\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "op-072->array-072\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "array-072->op-078\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "op-020\n",
+       "\n",
+       "\n",
+       "op-020\n",
+       "from_zarr\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "array-020\n",
+       "\n",
+       "\n",
+       "array-020\n",
+       "dy\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "op-020->array-020\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "op-076\n",
+       "\n",
+       "\n",
+       "op-076\n",
+       "__getitem__\n",
+       "tasks: 1\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "array-020->op-076\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "array-076\n",
+       "\n",
+       "\n",
+       "array-076\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "op-076->array-076\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "array-076->op-078\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "array-078\n",
+       "\n",
+       "\n",
+       "array-078\n",
+       "None\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "op-078->array-078\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "op-079\n",
+       "\n",
+       "\n",
+       "op-079\n",
+       "mean\n",
+       "tasks: 2\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "array-078->op-079\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "array-079\n",
+       "\n",
+       "\n",
+       "array-079\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "op-079->array-079\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "op-082\n",
+       "\n",
+       "\n",
+       "op-082\n",
+       "mean\n",
+       "tasks: 1\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "array-079->op-082\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "array-082\n",
+       "\n",
+       "\n",
+       "array-082\n",
+       "None\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "op-082->array-082\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "create-arrays\n",
+       "\n",
+       "\n",
+       "create-arrays\n",
+       "tasks: 7\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "arrays\n",
+       "\n",
+       "\n",
+       "arrays\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "create-arrays->arrays\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       ""
+      ],
+      "text/plain": [
+       ""
+      ]
+     },
+     "execution_count": 13,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "mean.data.visualize()"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 14,
+   "id": "f64e2ea9",
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "application/vnd.jupyter.widget-view+json": {
+       "model_id": "ea93bbb1ab8144d2b51fdadceab4a5f7",
+       "version_major": 2,
+       "version_minor": 0
+      },
+      "text/plain": [
+       "Output()"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    },
+    {
+     "data": {
+      "text/html": [
+       "
\n"
+      ],
+      "text/plain": []
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    },
+    {
+     "data": {
+      "text/html": [
+       "
\n",
+       "
\n"
+      ],
+      "text/plain": [
+       "\n"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    },
+    {
+     "data": {
+      "text/html": [
+       "
\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "
<xarray.DataArray (face: 1, j: 987, i: 1920)> Size: 15MB\n",
+       "array([[[-0.00935039,  0.05180967,  0.02368589, ...,  0.09296773,\n",
+       "         -0.02251309,  0.00811106],\n",
+       "        [-0.03833913,  0.05900023,  0.01459681, ..., -0.01437545,\n",
+       "          0.01901828,  0.02181335],\n",
+       "        [ 0.00342973,  0.01948303,  0.01160908, ...,  0.02988825,\n",
+       "         -0.0061075 , -0.01061161],\n",
+       "        ...,\n",
+       "        [ 0.03900664,  0.04362473,  0.00623131, ..., -0.03877337,\n",
+       "          0.00551559, -0.0664945 ],\n",
+       "        [-0.01084767, -0.00746532,  0.04994078, ...,  0.00983424,\n",
+       "         -0.00747178, -0.01885188],\n",
+       "        [ 0.00544092,  0.05520928, -0.0255991 , ...,  0.02842916,\n",
+       "          0.0289831 ,  0.01773723]]])\n",
+       "Dimensions without coordinates: face, j, i
"
+      ],
+      "text/plain": [
+       " Size: 15MB\n",
+       "array([[[-0.00935039,  0.05180967,  0.02368589, ...,  0.09296773,\n",
+       "         -0.02251309,  0.00811106],\n",
+       "        [-0.03833913,  0.05900023,  0.01459681, ..., -0.01437545,\n",
+       "          0.01901828,  0.02181335],\n",
+       "        [ 0.00342973,  0.01948303,  0.01160908, ...,  0.02988825,\n",
+       "         -0.0061075 , -0.01061161],\n",
+       "        ...,\n",
+       "        [ 0.03900664,  0.04362473,  0.00623131, ..., -0.03877337,\n",
+       "          0.00551559, -0.0664945 ],\n",
+       "        [-0.01084767, -0.00746532,  0.04994078, ...,  0.00983424,\n",
+       "         -0.00747178, -0.01885188],\n",
+       "        [ 0.00544092,  0.05520928, -0.0255991 , ...,  0.02842916,\n",
+       "          0.0289831 ,  0.01773723]]])\n",
+       "Dimensions without coordinates: face, j, i"
+      ]
+     },
+     "execution_count": 14,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "mean.compute(callbacks=[RichProgressBar()])"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "9572a661",
+   "metadata": {},
+   "outputs": [],
+   "source": []
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "Python 3 (ipykernel)",
+   "language": "python",
+   "name": "python3"
+  },
+  "language_info": {
+   "codemirror_mode": {
+    "name": "ipython",
+    "version": 3
+   },
+   "file_extension": ".py",
+   "mimetype": "text/x-python",
+   "name": "python",
+   "nbconvert_exporter": "python",
+   "pygments_lexer": "ipython3",
+   "version": "3.9.16"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 5
+}
diff --git a/examples/pangeo-2-quadratic-means.ipynb b/examples/pangeo-2-quadratic-means.ipynb
new file mode 100644
index 00000000..727e65df
--- /dev/null
+++ b/examples/pangeo-2-quadratic-means.ipynb
@@ -0,0 +1,2463 @@
+{
+ "cells": [
+  {
+   "cell_type": "markdown",
+   "id": "f575fe3a-48fc-4643-808c-89d4cd5d1c5d",
+   "metadata": {},
+   "source": [
+    "# Pangeo Quadratic Means Workload\n",
+    "\n",
+    "This is a notebook for exploring the example in https://github.com/pangeo-data/distributed-array-examples/issues/2.\n",
+    "\n",
+    "See also Tom Nicholas's [notebook](https://gist.github.com/TomNicholas/8366c917349b647d87860a20a257a3fb#file-benchmark-quadratic-ipynb) and Tom White's blog post, [Optimizing Cubed](https://medium.com/pangeo/optimizing-cubed-7a0b8f65f5b7)."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 1,
+   "id": "dba3adc2-d5b8-4c2d-9bc6-c434d4f573b9",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "import cubed\n",
+    "import cubed.random\n",
+    "from cubed.extensions.rich import RichProgressBar\n",
+    "import xarray as xr"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "f7238837-e6d7-4499-8df6-bd99b9ca26f6",
+   "metadata": {},
+   "source": [
+    "Initialization parameters for the workload"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 2,
+   "id": "1bb67eb5-e6ec-4254-b7e1-8a2318911038",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "t_length = 50\n",
+    "spec = cubed.Spec(allowed_mem=\"2GB\")"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "b51c5bd8-d27e-4699-8e3c-d650e5b66b72",
+   "metadata": {},
+   "source": [
+    "Create random data stored in Zarr."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 3,
+   "id": "f0a94b85-8146-4ffb-9d6c-4ccfc1da0575",
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "application/vnd.jupyter.widget-view+json": {
+       "model_id": "1a3d5ba8c0344fabb406b4a47484e68b",
+       "version_major": 2,
+       "version_minor": 0
+      },
+      "text/plain": [
+       "Output()"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    },
+    {
+     "data": {
+      "text/html": [
+       "
\n"
+      ],
+      "text/plain": []
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    },
+    {
+     "data": {
+      "text/html": [
+       "
\n",
+       "
\n"
+      ],
+      "text/plain": [
+       "\n"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    }
+   ],
+   "source": [
+    "u = cubed.random.random((t_length, 1, 987, 1920), chunks=(10, 1, -1, -1), spec=spec)\n",
+    "v = cubed.random.random((t_length, 1, 987, 1920), chunks=(10, 1, -1, -1), spec=spec)\n",
+    "arrays = [u, v]\n",
+    "paths = [f\"{spec.work_dir}/u_{t_length}.zarr\", f\"{spec.work_dir}/v_{t_length}.zarr\"]\n",
+    "cubed.store(arrays, paths, compute_arrays_in_parallel=True, callbacks=[RichProgressBar()])"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 4,
+   "id": "13c4e020-a968-4f4d-bfab-5d5d487dfb30",
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/html": [
+       "
\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "
<xarray.Dataset> Size: 2GB\n",
+       "Dimensions:  (time: 50, face: 1, j: 987, i: 1920)\n",
+       "Dimensions without coordinates: time, face, j, i\n",
+       "Data variables:\n",
+       "    anom_u   (time, face, j, i) float64 758MB cubed.Array<chunksize=(10, 1, 987, 1920)>\n",
+       "    anom_v   (time, face, j, i) float64 758MB cubed.Array<chunksize=(10, 1, 987, 1920)>
"
+      ],
+      "text/plain": [
+       " Size: 2GB\n",
+       "Dimensions:  (time: 50, face: 1, j: 987, i: 1920)\n",
+       "Dimensions without coordinates: time, face, j, i\n",
+       "Data variables:\n",
+       "    anom_u   (time, face, j, i) float64 758MB cubed.Array\n",
+       "    anom_v   (time, face, j, i) float64 758MB cubed.Array"
+      ]
+     },
+     "execution_count": 4,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "u = cubed.from_zarr(paths[0], spec=spec)\n",
+    "v = cubed.from_zarr(paths[1], spec=spec)\n",
+    "ds = xr.Dataset(\n",
+    "    dict(\n",
+    "        anom_u=([\"time\", \"face\", \"j\", \"i\"], u),\n",
+    "        anom_v=([\"time\", \"face\", \"j\", \"i\"], v),\n",
+    "    )\n",
+    ")\n",
+    "ds"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 5,
+   "id": "283d3c71-5271-499d-8743-ff489e4c5b90",
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/html": [
+       "
\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "
<xarray.Dataset> Size: 45MB\n",
+       "Dimensions:  (face: 1, j: 987, i: 1920)\n",
+       "Dimensions without coordinates: face, j, i\n",
+       "Data variables:\n",
+       "    anom_u   (face, j, i) float64 15MB cubed.Array<chunksize=(1, 987, 1920)>\n",
+       "    anom_v   (face, j, i) float64 15MB cubed.Array<chunksize=(1, 987, 1920)>\n",
+       "    uv       (face, j, i) float64 15MB cubed.Array<chunksize=(1, 987, 1920)>
"
+      ],
+      "text/plain": [
+       " Size: 45MB\n",
+       "Dimensions:  (face: 1, j: 987, i: 1920)\n",
+       "Dimensions without coordinates: face, j, i\n",
+       "Data variables:\n",
+       "    anom_u   (face, j, i) float64 15MB cubed.Array\n",
+       "    anom_v   (face, j, i) float64 15MB cubed.Array\n",
+       "    uv       (face, j, i) float64 15MB cubed.Array"
+      ]
+     },
+     "execution_count": 5,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "quad = ds**2\n",
+    "quad[\"uv\"] = ds.anom_u * ds.anom_v\n",
+    "result = quad.mean(\"time\", skipna=False)\n",
+    "result"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 6,
+   "id": "e97aa284",
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "image/svg+xml": [
+       "\n",
+       "\n",
+       "\n",
+       "num tasks: 33\n",
+       "max projected memory: 909.6 MB\n",
+       "total nbytes written: 2.5 GB\n",
+       "optimized: True\n",
+       "\n",
+       "\n",
+       "op-009\n",
+       "\n",
+       "\n",
+       "op-009\n",
+       "from_zarr\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "array-009\n",
+       "\n",
+       "\n",
+       "array-009\n",
+       "u\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "op-009->array-009\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "op-012\n",
+       "\n",
+       "\n",
+       "op-012\n",
+       "__pow__\n",
+       "tasks: 5\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "array-009->op-012\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "op-015\n",
+       "\n",
+       "\n",
+       "op-015\n",
+       "__mul__\n",
+       "tasks: 5\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "array-009->op-015\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "op-011\n",
+       "\n",
+       "\n",
+       "op-011\n",
+       "__pow__\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "array-011\n",
+       "\n",
+       "\n",
+       "array-011\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "op-011->array-011\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "array-011->op-012\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "array-012\n",
+       "\n",
+       "\n",
+       "array-012\n",
+       "values\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "op-012->array-012\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "op-016\n",
+       "\n",
+       "\n",
+       "op-016\n",
+       "mean\n",
+       "tasks: 2\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "array-012->op-016\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "array-016\n",
+       "\n",
+       "\n",
+       "array-016\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "op-016->array-016\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "op-019\n",
+       "\n",
+       "\n",
+       "op-019\n",
+       "mean\n",
+       "tasks: 1\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "array-016->op-019\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "array-019\n",
+       "\n",
+       "\n",
+       "array-019\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "op-019->array-019\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "op-010\n",
+       "\n",
+       "\n",
+       "op-010\n",
+       "from_zarr\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "array-010\n",
+       "\n",
+       "\n",
+       "array-010\n",
+       "v\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "op-010->array-010\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "op-014\n",
+       "\n",
+       "\n",
+       "op-014\n",
+       "__pow__\n",
+       "tasks: 5\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "array-010->op-014\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "array-010->op-015\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "op-013\n",
+       "\n",
+       "\n",
+       "op-013\n",
+       "__pow__\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "array-013\n",
+       "\n",
+       "\n",
+       "array-013\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "op-013->array-013\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "array-013->op-014\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "array-014\n",
+       "\n",
+       "\n",
+       "array-014\n",
+       "values\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "op-014->array-014\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "op-020\n",
+       "\n",
+       "\n",
+       "op-020\n",
+       "mean\n",
+       "tasks: 2\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "array-014->op-020\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "array-020\n",
+       "\n",
+       "\n",
+       "array-020\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "op-020->array-020\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "op-023\n",
+       "\n",
+       "\n",
+       "op-023\n",
+       "mean\n",
+       "tasks: 1\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "array-020->op-023\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "array-023\n",
+       "\n",
+       "\n",
+       "array-023\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "op-023->array-023\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "array-015\n",
+       "\n",
+       "\n",
+       "array-015\n",
+       "values\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "op-015->array-015\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "op-024\n",
+       "\n",
+       "\n",
+       "op-024\n",
+       "mean\n",
+       "tasks: 2\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "array-015->op-024\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "array-024\n",
+       "\n",
+       "\n",
+       "array-024\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "op-024->array-024\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "op-027\n",
+       "\n",
+       "\n",
+       "op-027\n",
+       "mean\n",
+       "tasks: 1\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "array-024->op-027\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "array-027\n",
+       "\n",
+       "\n",
+       "array-027\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "op-027->array-027\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "create-arrays\n",
+       "\n",
+       "\n",
+       "create-arrays\n",
+       "tasks: 9\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "arrays\n",
+       "\n",
+       "\n",
+       "arrays\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "create-arrays->arrays\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       ""
+      ],
+      "text/plain": [
+       ""
+      ]
+     },
+     "execution_count": 6,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "cubed.visualize(\n",
+    "    *(result[var].data for var in (\"anom_u\", \"anom_v\", \"uv\")),\n",
+    ")"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 7,
+   "id": "1700ce7b",
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "application/vnd.jupyter.widget-view+json": {
+       "model_id": "ea28a388169c4e47a463759111801687",
+       "version_major": 2,
+       "version_minor": 0
+      },
+      "text/plain": [
+       "Output()"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    },
+    {
+     "data": {
+      "text/html": [
+       "
\n"
+      ],
+      "text/plain": []
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    },
+    {
+     "data": {
+      "text/html": [
+       "
\n",
+       "
\n"
+      ],
+      "text/plain": [
+       "\n"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    },
+    {
+     "data": {
+      "text/html": [
+       "
\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "
<xarray.Dataset> Size: 45MB\n",
+       "Dimensions:  (face: 1, j: 987, i: 1920)\n",
+       "Dimensions without coordinates: face, j, i\n",
+       "Data variables:\n",
+       "    anom_u   (face, j, i) float64 15MB 0.339 0.3298 0.3399 ... 0.3794 0.2628\n",
+       "    anom_v   (face, j, i) float64 15MB 0.3566 0.3384 0.3565 ... 0.3379 0.3464\n",
+       "    uv       (face, j, i) float64 15MB 0.2717 0.2679 0.281 ... 0.2727 0.2272
"
+      ],
+      "text/plain": [
+       " Size: 45MB\n",
+       "Dimensions:  (face: 1, j: 987, i: 1920)\n",
+       "Dimensions without coordinates: face, j, i\n",
+       "Data variables:\n",
+       "    anom_u   (face, j, i) float64 15MB 0.339 0.3298 0.3399 ... 0.3794 0.2628\n",
+       "    anom_v   (face, j, i) float64 15MB 0.3566 0.3384 0.3565 ... 0.3379 0.3464\n",
+       "    uv       (face, j, i) float64 15MB 0.2717 0.2679 0.281 ... 0.2727 0.2272"
+      ]
+     },
+     "execution_count": 7,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "result.compute(callbacks=[RichProgressBar()])"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "b345817e",
+   "metadata": {},
+   "outputs": [],
+   "source": []
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "Python 3 (ipykernel)",
+   "language": "python",
+   "name": "python3"
+  },
+  "language_info": {
+   "codemirror_mode": {
+    "name": "ipython",
+    "version": 3
+   },
+   "file_extension": ".py",
+   "mimetype": "text/x-python",
+   "name": "python",
+   "nbconvert_exporter": "python",
+   "pygments_lexer": "ipython3",
+   "version": "3.9.16"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 5
+}
diff --git a/examples/pangeo-3-tem.ipynb b/examples/pangeo-3-tem.ipynb
new file mode 100644
index 00000000..7f011801
--- /dev/null
+++ b/examples/pangeo-3-tem.ipynb
@@ -0,0 +1,8318 @@
+{
+ "cells": [
+  {
+   "cell_type": "markdown",
+   "id": "f575fe3a-48fc-4643-808c-89d4cd5d1c5d",
+   "metadata": {},
+   "source": [
+    "# Pangeo Transformed Eulerian Mean Workload\n",
+    "\n",
+    "This is a notebook for exploring the example in https://github.com/pangeo-data/distributed-array-examples/issues/3.\n",
+    "\n",
+    "See also https://github.com/dcherian/ncar-challenge-suite/blob/main/tem.ipynb for the original, and the Cubed tracking issue at https://github.com/tomwhite/cubed/issues/145"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 1,
+   "id": "dba3adc2-d5b8-4c2d-9bc6-c434d4f573b9",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "import cubed\n",
+    "import cubed.random\n",
+    "from cubed.extensions.rich import RichProgressBar\n",
+    "import pandas as pd\n",
+    "import xarray as xr"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "f7238837-e6d7-4499-8df6-bd99b9ca26f6",
+   "metadata": {},
+   "source": [
+    "Initialization parameters for the workload"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 2,
+   "id": "1bb67eb5-e6ec-4254-b7e1-8a2318911038",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "dims = (\"time\", \"level\", \"lat\", \"lon\")\n",
+    "# nyears is number of years, adjust to make bigger, \n",
+    "# full dataset is 60-ish years.\n",
+    "nyears = 1\n",
+    "#shape = (nyears * 365 * 24, 37, 721, 1440)\n",
+    "shape = (nyears * 31 * 24, 37, 72, 144)  # make smaller while exploring locally\n",
+    "chunks = (24, 1, -1, -1)\n",
+    "spec = cubed.Spec(allowed_mem=\"2GB\")"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "b51c5bd8-d27e-4699-8e3c-d650e5b66b72",
+   "metadata": {},
+   "source": [
+    "Create random data stored in Zarr."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 3,
+   "id": "f0a94b85-8146-4ffb-9d6c-4ccfc1da0575",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "u = cubed.random.random(shape, chunks=chunks, spec=spec)\n",
+    "v = cubed.random.random(shape, chunks=chunks, spec=spec)\n",
+    "w = cubed.random.random(shape, chunks=chunks, spec=spec)\n",
+    "t = cubed.random.random(shape, chunks=chunks, spec=spec)\n",
+    "arrays = [u, v, w, t]\n",
+    "paths = [\n",
+    "    f\"{spec.work_dir}/u_{nyears}.zarr\",\n",
+    "    f\"{spec.work_dir}/v_{nyears}.zarr\",\n",
+    "    f\"{spec.work_dir}/w_{nyears}.zarr\",\n",
+    "    f\"{spec.work_dir}/t_{nyears}.zarr\",\n",
+    "]"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 4,
+   "id": "20d4b17b",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# Comment in below to save then load Zarr from filesystem\n",
+    "# cubed.store(arrays, paths, compute_arrays_in_parallel=True, callbacks=[RichProgressBar()])"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 5,
+   "id": "9ede34e3",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# u = cubed.from_zarr(paths[0], spec=spec)\n",
+    "# v = cubed.from_zarr(paths[1], spec=spec)\n",
+    "# w = cubed.from_zarr(paths[2], spec=spec)\n",
+    "# t = cubed.from_zarr(paths[3], spec=spec)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 6,
+   "id": "13c4e020-a968-4f4d-bfab-5d5d487dfb30",
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/html": [
+       "
\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "
<xarray.Dataset> Size: 9GB\n",
+       "Dimensions:  (time: 744, level: 37, lat: 72, lon: 144)\n",
+       "Coordinates:\n",
+       "  * time     (time) datetime64[ns] 6kB 2001-01-01 ... 2001-01-31T23:00:00\n",
+       "Dimensions without coordinates: level, lat, lon\n",
+       "Data variables:\n",
+       "    U        (time, level, lat, lon) float64 2GB cubed.Array<chunksize=(24, 1, 72, 144)>\n",
+       "    V        (time, level, lat, lon) float64 2GB cubed.Array<chunksize=(24, 1, 72, 144)>\n",
+       "    W        (time, level, lat, lon) float64 2GB cubed.Array<chunksize=(24, 1, 72, 144)>\n",
+       "    T        (time, level, lat, lon) float64 2GB cubed.Array<chunksize=(24, 1, 72, 144)>
"
+      ],
+      "text/plain": [
+       " Size: 9GB\n",
+       "Dimensions:  (time: 744, level: 37, lat: 72, lon: 144)\n",
+       "Coordinates:\n",
+       "  * time     (time) datetime64[ns] 6kB 2001-01-01 ... 2001-01-31T23:00:00\n",
+       "Dimensions without coordinates: level, lat, lon\n",
+       "Data variables:\n",
+       "    U        (time, level, lat, lon) float64 2GB cubed.Array\n",
+       "    V        (time, level, lat, lon) float64 2GB cubed.Array\n",
+       "    W        (time, level, lat, lon) float64 2GB cubed.Array\n",
+       "    T        (time, level, lat, lon) float64 2GB cubed.Array"
+      ]
+     },
+     "execution_count": 6,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "ds = xr.Dataset(\n",
+    "    {\n",
+    "        \"U\": (dims, u),\n",
+    "        \"V\": (dims, v),\n",
+    "        \"W\": (dims, w),\n",
+    "        \"T\": (dims, t),\n",
+    "    },\n",
+    "    coords={\"time\": pd.date_range(\"2001-01-01\", periods=shape[0], freq=\"H\")},\n",
+    ")\n",
+    "ds"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 7,
+   "id": "283d3c71-5271-499d-8743-ff489e4c5b90",
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/html": [
+       "
\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "
<xarray.Dataset> Size: 63MB\n",
+       "Dimensions:  (time: 744, level: 37, lat: 72)\n",
+       "Coordinates:\n",
+       "  * time     (time) datetime64[ns] 6kB 2001-01-01 ... 2001-01-31T23:00:00\n",
+       "Dimensions without coordinates: level, lat\n",
+       "Data variables:\n",
+       "    U        (time, level, lat) float64 16MB cubed.Array<chunksize=(24, 1, 72)>\n",
+       "    V        (time, level, lat) float64 16MB cubed.Array<chunksize=(24, 1, 72)>\n",
+       "    W        (time, level, lat) float64 16MB cubed.Array<chunksize=(24, 1, 72)>\n",
+       "    T        (time, level, lat) float64 16MB cubed.Array<chunksize=(24, 1, 72)>
"
+      ],
+      "text/plain": [
+       " Size: 63MB\n",
+       "Dimensions:  (time: 744, level: 37, lat: 72)\n",
+       "Coordinates:\n",
+       "  * time     (time) datetime64[ns] 6kB 2001-01-01 ... 2001-01-31T23:00:00\n",
+       "Dimensions without coordinates: level, lat\n",
+       "Data variables:\n",
+       "    U        (time, level, lat) float64 16MB cubed.Array\n",
+       "    V        (time, level, lat) float64 16MB cubed.Array\n",
+       "    W        (time, level, lat) float64 16MB cubed.Array\n",
+       "    T        (time, level, lat) float64 16MB cubed.Array"
+      ]
+     },
+     "execution_count": 7,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "# Note: we actually want skipna=True, but this isn't implemented in xarray yet\n",
+    "# see https://github.com/pydata/xarray/issues/7243\n",
+    "zonal_means = ds.mean(\"lon\", skipna=False)\n",
+    "zonal_means"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 8,
+   "id": "fcfc18e3",
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/html": [
+       "
\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "
<xarray.Dataset> Size: 16GB\n",
+       "Dimensions:  (time: 744, level: 37, lat: 72, lon: 144)\n",
+       "Coordinates:\n",
+       "  * time     (time) datetime64[ns] 6kB 2001-01-01 ... 2001-01-31T23:00:00\n",
+       "Dimensions without coordinates: level, lat, lon\n",
+       "Data variables:\n",
+       "    U        (time, level, lat, lon) float64 2GB cubed.Array<chunksize=(24, 1, 72, 144)>\n",
+       "    V        (time, level, lat, lon) float64 2GB cubed.Array<chunksize=(24, 1, 72, 144)>\n",
+       "    W        (time, level, lat, lon) float64 2GB cubed.Array<chunksize=(24, 1, 72, 144)>\n",
+       "    T        (time, level, lat, lon) float64 2GB cubed.Array<chunksize=(24, 1, 72, 144)>\n",
+       "    uv       (time, level, lat, lon) float64 2GB cubed.Array<chunksize=(24, 1, 72, 144)>\n",
+       "    vt       (time, level, lat, lon) float64 2GB cubed.Array<chunksize=(24, 1, 72, 144)>\n",
+       "    uw       (time, level, lat, lon) float64 2GB cubed.Array<chunksize=(24, 1, 72, 144)>
"
+      ],
+      "text/plain": [
+       " Size: 16GB\n",
+       "Dimensions:  (time: 744, level: 37, lat: 72, lon: 144)\n",
+       "Coordinates:\n",
+       "  * time     (time) datetime64[ns] 6kB 2001-01-01 ... 2001-01-31T23:00:00\n",
+       "Dimensions without coordinates: level, lat, lon\n",
+       "Data variables:\n",
+       "    U        (time, level, lat, lon) float64 2GB cubed.Array\n",
+       "    V        (time, level, lat, lon) float64 2GB cubed.Array\n",
+       "    W        (time, level, lat, lon) float64 2GB cubed.Array\n",
+       "    T        (time, level, lat, lon) float64 2GB cubed.Array\n",
+       "    uv       (time, level, lat, lon) float64 2GB cubed.Array\n",
+       "    vt       (time, level, lat, lon) float64 2GB cubed.Array\n",
+       "    uw       (time, level, lat, lon) float64 2GB cubed.Array"
+      ]
+     },
+     "execution_count": 8,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "anomaly = ds - zonal_means\n",
+    "anomaly['uv'] = anomaly.U*anomaly.V\n",
+    "anomaly['vt'] = anomaly.V*anomaly.T\n",
+    "anomaly['uw'] = anomaly.U*anomaly.W\n",
+    "anomaly"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 9,
+   "id": "cce3f02a",
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/html": [
+       "
\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "
<xarray.Dataset> Size: 111MB\n",
+       "Dimensions:  (time: 744, level: 37, lat: 72)\n",
+       "Coordinates:\n",
+       "  * time     (time) datetime64[ns] 6kB 2001-01-01 ... 2001-01-31T23:00:00\n",
+       "Dimensions without coordinates: level, lat\n",
+       "Data variables:\n",
+       "    U        (time, level, lat) float64 16MB cubed.Array<chunksize=(24, 1, 72)>\n",
+       "    V        (time, level, lat) float64 16MB cubed.Array<chunksize=(24, 1, 72)>\n",
+       "    W        (time, level, lat) float64 16MB cubed.Array<chunksize=(24, 1, 72)>\n",
+       "    T        (time, level, lat) float64 16MB cubed.Array<chunksize=(24, 1, 72)>\n",
+       "    uv       (time, level, lat) float64 16MB cubed.Array<chunksize=(24, 1, 72)>\n",
+       "    vt       (time, level, lat) float64 16MB cubed.Array<chunksize=(24, 1, 72)>\n",
+       "    uw       (time, level, lat) float64 16MB cubed.Array<chunksize=(24, 1, 72)>
"
+      ],
+      "text/plain": [
+       " Size: 111MB\n",
+       "Dimensions:  (time: 744, level: 37, lat: 72)\n",
+       "Coordinates:\n",
+       "  * time     (time) datetime64[ns] 6kB 2001-01-01 ... 2001-01-31T23:00:00\n",
+       "Dimensions without coordinates: level, lat\n",
+       "Data variables:\n",
+       "    U        (time, level, lat) float64 16MB cubed.Array\n",
+       "    V        (time, level, lat) float64 16MB cubed.Array\n",
+       "    W        (time, level, lat) float64 16MB cubed.Array\n",
+       "    T        (time, level, lat) float64 16MB cubed.Array\n",
+       "    uv       (time, level, lat) float64 16MB cubed.Array\n",
+       "    vt       (time, level, lat) float64 16MB cubed.Array\n",
+       "    uw       (time, level, lat) float64 16MB cubed.Array"
+      ]
+     },
+     "execution_count": 9,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "temdiags = zonal_means.merge(anomaly[['uv','vt','uw']].mean(\"lon\", skipna=False))  # actually want skipna=True\n",
+    "temdiags"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 10,
+   "id": "6734ec3f",
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/html": [
+       "
\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "
<xarray.Dataset> Size: 5MB\n",
+       "Dimensions:  (time: 31, level: 37, lat: 72)\n",
+       "Coordinates:\n",
+       "  * time     (time) datetime64[ns] 248B 2001-01-01 2001-01-02 ... 2001-01-31\n",
+       "Dimensions without coordinates: level, lat\n",
+       "Data variables:\n",
+       "    U        (time, level, lat) float64 661kB cubed.Array<chunksize=(1, 1, 72)>\n",
+       "    V        (time, level, lat) float64 661kB cubed.Array<chunksize=(1, 1, 72)>\n",
+       "    W        (time, level, lat) float64 661kB cubed.Array<chunksize=(1, 1, 72)>\n",
+       "    T        (time, level, lat) float64 661kB cubed.Array<chunksize=(1, 1, 72)>\n",
+       "    uv       (time, level, lat) float64 661kB cubed.Array<chunksize=(1, 1, 72)>\n",
+       "    vt       (time, level, lat) float64 661kB cubed.Array<chunksize=(1, 1, 72)>\n",
+       "    uw       (time, level, lat) float64 661kB cubed.Array<chunksize=(1, 1, 72)>
"
+      ],
+      "text/plain": [
+       " Size: 5MB\n",
+       "Dimensions:  (time: 31, level: 37, lat: 72)\n",
+       "Coordinates:\n",
+       "  * time     (time) datetime64[ns] 248B 2001-01-01 2001-01-02 ... 2001-01-31\n",
+       "Dimensions without coordinates: level, lat\n",
+       "Data variables:\n",
+       "    U        (time, level, lat) float64 661kB cubed.Array\n",
+       "    V        (time, level, lat) float64 661kB cubed.Array\n",
+       "    W        (time, level, lat) float64 661kB cubed.Array\n",
+       "    T        (time, level, lat) float64 661kB cubed.Array\n",
+       "    uv       (time, level, lat) float64 661kB cubed.Array\n",
+       "    vt       (time, level, lat) float64 661kB cubed.Array\n",
+       "    uw       (time, level, lat) float64 661kB cubed.Array"
+      ]
+     },
+     "execution_count": 10,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "# note method=\"blockwise\" uses flox\n",
+    "temdiags = temdiags.resample(time=\"D\").mean(method=\"blockwise\", skipna=False)  # actually want skipna=True\n",
+    "temdiags"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 11,
+   "id": "e97aa284",
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "image/svg+xml": [
+       "\n",
+       "\n",
+       "\n",
+       "num tasks: 11480\n",
+       "max projected memory: 11.9 MB\n",
+       "total nbytes written: 113.0 MB\n",
+       "optimized: True\n",
+       "\n",
+       "\n",
+       "op-015\n",
+       "\n",
+       "\n",
+       "op-015\n",
+       "mean\n",
+       "tasks: 1147\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "array-015\n",
+       "\n",
+       "\n",
+       "array-015\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "op-015->array-015\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "op-065\n",
+       "\n",
+       "\n",
+       "op-065\n",
+       "permute_dims\n",
+       "tasks: 1147\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "array-015->op-065\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "op-067\n",
+       "\n",
+       "\n",
+       "op-067\n",
+       "permute_dims\n",
+       "tasks: 1147\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "array-015->op-067\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "op-018\n",
+       "\n",
+       "\n",
+       "op-018\n",
+       "mean\n",
+       "tasks: 1147\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "array-018\n",
+       "\n",
+       "\n",
+       "array-018\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "op-018->array-018\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "array-018->op-065\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "op-066\n",
+       "\n",
+       "\n",
+       "op-066\n",
+       "permute_dims\n",
+       "tasks: 1147\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "array-018->op-066\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "array-065\n",
+       "\n",
+       "\n",
+       "array-065\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "op-065->array-065\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "op-093\n",
+       "\n",
+       "\n",
+       "op-093\n",
+       "permute_dims\n",
+       "tasks: 1147\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "array-065->op-093\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "array-093\n",
+       "\n",
+       "\n",
+       "array-093\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "op-093->array-093\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "op-024\n",
+       "\n",
+       "\n",
+       "op-024\n",
+       "mean\n",
+       "tasks: 1147\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "array-024\n",
+       "\n",
+       "\n",
+       "array-024\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "op-024->array-024\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "array-024->op-066\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "array-066\n",
+       "\n",
+       "\n",
+       "array-066\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "op-066->array-066\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "op-094\n",
+       "\n",
+       "\n",
+       "op-094\n",
+       "permute_dims\n",
+       "tasks: 1147\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "array-066->op-094\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "array-094\n",
+       "\n",
+       "\n",
+       "array-094\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "op-094->array-094\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "op-021\n",
+       "\n",
+       "\n",
+       "op-021\n",
+       "mean\n",
+       "tasks: 1147\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "array-021\n",
+       "\n",
+       "\n",
+       "array-021\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "op-021->array-021\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "array-021->op-067\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "array-067\n",
+       "\n",
+       "\n",
+       "array-067\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "op-067->array-067\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "op-095\n",
+       "\n",
+       "\n",
+       "op-095\n",
+       "permute_dims\n",
+       "tasks: 1147\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "array-067->op-095\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "array-095\n",
+       "\n",
+       "\n",
+       "array-095\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "op-095->array-095\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "create-arrays\n",
+       "\n",
+       "\n",
+       "create-arrays\n",
+       "tasks: 10\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "arrays\n",
+       "\n",
+       "\n",
+       "arrays\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "create-arrays->arrays\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       ""
+      ],
+      "text/plain": [
+       ""
+      ]
+     },
+     "execution_count": 11,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "cubed.visualize(\n",
+    "    *(temdiags[var].data for var in (\"uv\", \"vt\", \"uw\")),\n",
+    ")"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 12,
+   "id": "1700ce7b",
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "application/vnd.jupyter.widget-view+json": {
+       "model_id": "f9e290547e7c47c8b5576dbfb40ce1e5",
+       "version_major": 2,
+       "version_minor": 0
+      },
+      "text/plain": [
+       "Output()"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    },
+    {
+     "data": {
+      "text/html": [
+       "
\n"
+      ],
+      "text/plain": []
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    },
+    {
+     "data": {
+      "text/html": [
+       "
\n",
+       "
\n"
+      ],
+      "text/plain": [
+       "\n"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    },
+    {
+     "data": {
+      "text/html": [
+       "
\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "
<xarray.Dataset> Size: 5MB\n",
+       "Dimensions:  (time: 31, level: 37, lat: 72)\n",
+       "Coordinates:\n",
+       "  * time     (time) datetime64[ns] 248B 2001-01-01 2001-01-02 ... 2001-01-31\n",
+       "Dimensions without coordinates: level, lat\n",
+       "Data variables:\n",
+       "    U        (time, level, lat) float64 661kB 0.502 0.5039 0.504 ... 0.5 0.4956\n",
+       "    V        (time, level, lat) float64 661kB 0.4953 0.4931 ... 0.4944 0.4976\n",
+       "    W        (time, level, lat) float64 661kB 0.5017 0.4984 ... 0.5083 0.5015\n",
+       "    T        (time, level, lat) float64 661kB 0.4949 0.503 ... 0.5024 0.487\n",
+       "    uv       (time, level, lat) float64 661kB 0.00129 0.0009764 ... 0.0003541\n",
+       "    vt       (time, level, lat) float64 661kB -0.002118 -0.001934 ... 0.001608\n",
+       "    uw       (time, level, lat) float64 661kB 0.0007928 -0.00245 ... -0.002159
"
+      ],
+      "text/plain": [
+       " Size: 5MB\n",
+       "Dimensions:  (time: 31, level: 37, lat: 72)\n",
+       "Coordinates:\n",
+       "  * time     (time) datetime64[ns] 248B 2001-01-01 2001-01-02 ... 2001-01-31\n",
+       "Dimensions without coordinates: level, lat\n",
+       "Data variables:\n",
+       "    U        (time, level, lat) float64 661kB 0.502 0.5039 0.504 ... 0.5 0.4956\n",
+       "    V        (time, level, lat) float64 661kB 0.4953 0.4931 ... 0.4944 0.4976\n",
+       "    W        (time, level, lat) float64 661kB 0.5017 0.4984 ... 0.5083 0.5015\n",
+       "    T        (time, level, lat) float64 661kB 0.4949 0.503 ... 0.5024 0.487\n",
+       "    uv       (time, level, lat) float64 661kB 0.00129 0.0009764 ... 0.0003541\n",
+       "    vt       (time, level, lat) float64 661kB -0.002118 -0.001934 ... 0.001608\n",
+       "    uw       (time, level, lat) float64 661kB 0.0007928 -0.00245 ... -0.002159"
+      ]
+     },
+     "execution_count": 12,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "temdiags.compute(callbacks=[RichProgressBar()])"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "b345817e",
+   "metadata": {},
+   "outputs": [],
+   "source": []
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "Python 3 (ipykernel)",
+   "language": "python",
+   "name": "python3"
+  },
+  "language_info": {
+   "codemirror_mode": {
+    "name": "ipython",
+    "version": 3
+   },
+   "file_extension": ".py",
+   "mimetype": "text/x-python",
+   "name": "python",
+   "nbconvert_exporter": "python",
+   "pygments_lexer": "ipython3",
+   "version": "3.9.16"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 5
+}
diff --git a/examples/pangeo-4-climatological-anomalies.ipynb b/examples/pangeo-4-climatological-anomalies.ipynb
new file mode 100644
index 00000000..5978de07
--- /dev/null
+++ b/examples/pangeo-4-climatological-anomalies.ipynb
@@ -0,0 +1,2493 @@
+{
+ "cells": [
+  {
+   "cell_type": "markdown",
+   "id": "f575fe3a-48fc-4643-808c-89d4cd5d1c5d",
+   "metadata": {},
+   "source": [
+    "# Pangeo Climatological Anomalies Workload\n",
+    "\n",
+    "This is a notebook for exploring the example in https://github.com/pangeo-data/distributed-array-examples/issues/4.\n",
+    "\n",
+    "Also includes code from https://github.com/cubed-dev/cubed/issues/223"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 1,
+   "id": "dba3adc2-d5b8-4c2d-9bc6-c434d4f573b9",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "from datetime import datetime, timedelta\n",
+    "import cubed\n",
+    "import cubed.random\n",
+    "from cubed.extensions.rich import RichProgressBar\n",
+    "import numpy as np\n",
+    "import xarray as xr"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "f7238837-e6d7-4499-8df6-bd99b9ca26f6",
+   "metadata": {},
+   "source": [
+    "Initialization parameters for the workload"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 2,
+   "id": "1bb67eb5-e6ec-4254-b7e1-8a2318911038",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "t_length = 372\n",
+    "time = np.arange(datetime(1979,1,1), datetime(2022,1,1), timedelta(hours=1)).astype('datetime64[ns]')\n",
+    "lat = np.linspace(-90.0, 90.0, 721)[::-1].astype(np.float32)\n",
+    "lon = np.linspace(0.0, 359.8, 1440).astype(np.float32)\n",
+    "spec = cubed.Spec(allowed_mem=\"3GB\")  # TODO: can we reduce to 2GB?"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "b51c5bd8-d27e-4699-8e3c-d650e5b66b72",
+   "metadata": {},
+   "source": [
+    "Create random data stored in Zarr."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 3,
+   "id": "f0a94b85-8146-4ffb-9d6c-4ccfc1da0575",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "asn = cubed.array_api.astype(cubed.random.random((t_length, 721, 1440), chunks=(31, -1, -1), spec=spec), np.float32)\n",
+    "arrays = [asn]\n",
+    "paths = [\n",
+    "    f\"{spec.work_dir}/asn_{t_length}.zarr\",\n",
+    "]"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 4,
+   "id": "20d4b17b",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# Comment in below to save then load Zarr from filesystem\n",
+    "# cubed.store(arrays, paths, compute_arrays_in_parallel=True, callbacks=[RichProgressBar()])"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 5,
+   "id": "9ede34e3",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# asn = cubed.from_zarr(paths[0], spec=spec)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 6,
+   "id": "13c4e020-a968-4f4d-bfab-5d5d487dfb30",
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/html": [
+       "
\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "
<xarray.Dataset> Size: 2GB\n",
+       "Dimensions:    (time: 372, latitude: 721, longitude: 1440)\n",
+       "Coordinates:\n",
+       "  * time       (time) datetime64[ns] 3kB 1979-01-01 ... 1979-01-16T11:00:00\n",
+       "  * latitude   (latitude) float32 3kB 90.0 89.75 89.5 ... -89.5 -89.75 -90.0\n",
+       "  * longitude  (longitude) float32 6kB 0.0 0.25 0.5001 ... 359.3 359.5 359.8\n",
+       "Data variables:\n",
+       "    asn        (time, latitude, longitude) float32 2GB cubed.Array<chunksize=(31, 721, 1440)>
"
+      ],
+      "text/plain": [
+       " Size: 2GB\n",
+       "Dimensions:    (time: 372, latitude: 721, longitude: 1440)\n",
+       "Coordinates:\n",
+       "  * time       (time) datetime64[ns] 3kB 1979-01-01 ... 1979-01-16T11:00:00\n",
+       "  * latitude   (latitude) float32 3kB 90.0 89.75 89.5 ... -89.5 -89.75 -90.0\n",
+       "  * longitude  (longitude) float32 6kB 0.0 0.25 0.5001 ... 359.3 359.5 359.8\n",
+       "Data variables:\n",
+       "    asn        (time, latitude, longitude) float32 2GB cubed.Array"
+      ]
+     },
+     "execution_count": 6,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "ds = xr.Dataset(\n",
+    "    {\n",
+    "        \"asn\": (['time', 'latitude', 'longitude'], asn),\n",
+    "    },\n",
+    "    coords={'time': time[:t_length], 'latitude': lat, 'longitude': lon},\n",
+    ")\n",
+    "ds"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 7,
+   "id": "1065a3d0",
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "image/svg+xml": [
+       "\n",
+       "\n",
+       "\n",
+       "num tasks: 13\n",
+       "max projected memory: 1.0 GB\n",
+       "total nbytes written: 1.5 GB\n",
+       "optimized: True\n",
+       "\n",
+       "\n",
+       "op-001\n",
+       "\n",
+       "\n",
+       "op-001\n",
+       "random\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "array-001\n",
+       "\n",
+       "\n",
+       "array-001\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "op-001->array-001\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "op-004\n",
+       "\n",
+       "\n",
+       "op-004\n",
+       "astype\n",
+       "tasks: 12\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "array-001->op-004\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "op-002\n",
+       "\n",
+       "\n",
+       "op-002\n",
+       "random\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "array-002\n",
+       "\n",
+       "\n",
+       "array-002\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "op-002->array-002\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "array-002->op-004\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "array-004\n",
+       "\n",
+       "\n",
+       "array-004\n",
+       "asn\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "op-004->array-004\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "create-arrays\n",
+       "\n",
+       "\n",
+       "create-arrays\n",
+       "tasks: 1\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "arrays\n",
+       "\n",
+       "\n",
+       "arrays\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "create-arrays->arrays\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       ""
+      ],
+      "text/plain": [
+       ""
+      ]
+     },
+     "execution_count": 7,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "ds.asn.data.visualize(show_hidden=True)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 8,
+   "id": "283d3c71-5271-499d-8743-ff489e4c5b90",
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/html": [
+       "
\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "
<xarray.Dataset> Size: 66MB\n",
+       "Dimensions:    (dayofyear: 16, latitude: 721, longitude: 1440)\n",
+       "Coordinates:\n",
+       "  * latitude   (latitude) float32 3kB 90.0 89.75 89.5 ... -89.5 -89.75 -90.0\n",
+       "  * longitude  (longitude) float32 6kB 0.0 0.25 0.5001 ... 359.3 359.5 359.8\n",
+       "  * dayofyear  (dayofyear) int64 128B 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16\n",
+       "Data variables:\n",
+       "    asn        (dayofyear, latitude, longitude) float32 66MB cubed.Array<chunksize=(16, 721, 1440)>
"
+      ],
+      "text/plain": [
+       " Size: 66MB\n",
+       "Dimensions:    (dayofyear: 16, latitude: 721, longitude: 1440)\n",
+       "Coordinates:\n",
+       "  * latitude   (latitude) float32 3kB 90.0 89.75 89.5 ... -89.5 -89.75 -90.0\n",
+       "  * longitude  (longitude) float32 6kB 0.0 0.25 0.5001 ... 359.3 359.5 359.8\n",
+       "  * dayofyear  (dayofyear) int64 128B 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16\n",
+       "Data variables:\n",
+       "    asn        (dayofyear, latitude, longitude) float32 66MB cubed.Array"
+      ]
+     },
+     "execution_count": 8,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "# Note: we actually want skipna=True, but this isn't implemented in xarray yet\n",
+    "# see https://github.com/pydata/xarray/issues/7243\n",
+    "mean = ds.groupby(\"time.dayofyear\").mean(method=\"map-reduce\", skipna=False)\n",
+    "mean"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 9,
+   "id": "fcfc18e3",
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/html": [
+       "
\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "
<xarray.Dataset> Size: 2GB\n",
+       "Dimensions:    (time: 372, latitude: 721, longitude: 1440)\n",
+       "Coordinates:\n",
+       "  * time       (time) datetime64[ns] 3kB 1979-01-01 ... 1979-01-16T11:00:00\n",
+       "  * latitude   (latitude) float32 3kB 90.0 89.75 89.5 ... -89.5 -89.75 -90.0\n",
+       "  * longitude  (longitude) float32 6kB 0.0 0.25 0.5001 ... 359.3 359.5 359.8\n",
+       "    dayofyear  (time) int64 3kB 1 1 1 1 1 1 1 1 1 ... 16 16 16 16 16 16 16 16 16\n",
+       "Data variables:\n",
+       "    asn        (time, latitude, longitude) float32 2GB cubed.Array<chunksize=(16, 721, 1440)>
"
+      ],
+      "text/plain": [
+       " Size: 2GB\n",
+       "Dimensions:    (time: 372, latitude: 721, longitude: 1440)\n",
+       "Coordinates:\n",
+       "  * time       (time) datetime64[ns] 3kB 1979-01-01 ... 1979-01-16T11:00:00\n",
+       "  * latitude   (latitude) float32 3kB 90.0 89.75 89.5 ... -89.5 -89.75 -90.0\n",
+       "  * longitude  (longitude) float32 6kB 0.0 0.25 0.5001 ... 359.3 359.5 359.8\n",
+       "    dayofyear  (time) int64 3kB 1 1 1 1 1 1 1 1 1 ... 16 16 16 16 16 16 16 16 16\n",
+       "Data variables:\n",
+       "    asn        (time, latitude, longitude) float32 2GB cubed.Array"
+      ]
+     },
+     "execution_count": 9,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "anomaly = ds - mean.sel(dayofyear=ds.time.dt.dayofyear)\n",
+    "anomaly"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 10,
+   "id": "33e5b223",
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "image/svg+xml": [
+       "\n",
+       "\n",
+       "\n",
+       "num tasks: 100\n",
+       "max projected memory: 3.0 GB\n",
+       "total nbytes written: 14.0 GB\n",
+       "optimized: True\n",
+       "\n",
+       "\n",
+       "op-004\n",
+       "\n",
+       "\n",
+       "op-004\n",
+       "astype\n",
+       "tasks: 12\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "array-004\n",
+       "\n",
+       "\n",
+       "array-004\n",
+       "asn\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "op-004->array-004\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "op-029\n",
+       "\n",
+       "\n",
+       "op-029\n",
+       "__sub__\n",
+       "tasks: 3\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "array-004->op-029\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "array-029\n",
+       "\n",
+       "\n",
+       "array-029\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "op-029->array-029\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "op-030\n",
+       "\n",
+       "\n",
+       "op-030\n",
+       "__sub__\n",
+       "tasks: 24\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "array-029->op-030\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "op-012\n",
+       "\n",
+       "\n",
+       "op-012\n",
+       "groupby_reduction\n",
+       "tasks: 12\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "array-012\n",
+       "\n",
+       "\n",
+       "array-012\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "op-012->array-012\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "op-013\n",
+       "\n",
+       "\n",
+       "op-013\n",
+       "groupby_reduction\n",
+       "tasks: 6\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "array-012->op-013\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "array-013\n",
+       "\n",
+       "\n",
+       "array-013\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "op-013->array-013\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "op-014\n",
+       "\n",
+       "\n",
+       "op-014\n",
+       "groupby_reduction\n",
+       "tasks: 3\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "array-013->op-014\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "array-014\n",
+       "\n",
+       "\n",
+       "array-014\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "op-014->array-014\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "op-015\n",
+       "\n",
+       "\n",
+       "op-015\n",
+       "groupby_reduction\n",
+       "tasks: 2\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "array-014->op-015\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "array-015\n",
+       "\n",
+       "\n",
+       "array-015\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "op-015->array-015\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "op-019\n",
+       "\n",
+       "\n",
+       "op-019\n",
+       "permute_dims\n",
+       "tasks: 1\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "array-015->op-019\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "array-019\n",
+       "\n",
+       "\n",
+       "array-019\n",
+       "value\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "op-019->array-019\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "op-022\n",
+       "\n",
+       "\n",
+       "op-022\n",
+       "__getitem__\n",
+       "tasks: 1\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "array-019->op-022\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "array-022\n",
+       "\n",
+       "\n",
+       "array-022\n",
+       "value\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "op-022->array-022\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "op-025\n",
+       "\n",
+       "\n",
+       "op-025\n",
+       "__getitem__\n",
+       "tasks: 1\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "array-022->op-025\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "array-025\n",
+       "\n",
+       "\n",
+       "array-025\n",
+       "value\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "op-025->array-025\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "op-028\n",
+       "\n",
+       "\n",
+       "op-028\n",
+       "__getitem__\n",
+       "tasks: 24\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "array-025->op-028\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "array-028\n",
+       "\n",
+       "\n",
+       "array-028\n",
+       "other_data\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "op-028->array-028\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "array-028->op-030\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "array-030\n",
+       "\n",
+       "\n",
+       "array-030\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "op-030->array-030\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "create-arrays\n",
+       "\n",
+       "\n",
+       "create-arrays\n",
+       "tasks: 11\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "arrays\n",
+       "\n",
+       "\n",
+       "arrays\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       "create-arrays->arrays\n",
+       "\n",
+       "\n",
+       "\n",
+       "\n",
+       ""
+      ],
+      "text/plain": [
+       ""
+      ]
+     },
+     "execution_count": 10,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "cubed.visualize(\n",
+    "    *(anomaly[var].data for var in [\"asn\"]),\n",
+    ")"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 11,
+   "id": "c7d3e127",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "! rm -rf anomaly.zarr"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 12,
+   "id": "b345817e",
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "application/vnd.jupyter.widget-view+json": {
+       "model_id": "b6babab972b4483aa5c88ee6c29fe709",
+       "version_major": 2,
+       "version_minor": 0
+      },
+      "text/plain": [
+       "Output()"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    },
+    {
+     "data": {
+      "text/html": [
+       "
\n"
+      ],
+      "text/plain": []
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    },
+    {
+     "data": {
+      "text/html": [
+       "
\n",
+       "
\n"
+      ],
+      "text/plain": [
+       "\n"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    },
+    {
+     "data": {
+      "text/plain": [
+       ""
+      ]
+     },
+     "execution_count": 12,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "anomaly.to_zarr(\"anomaly.zarr\", chunkmanager_store_kwargs=dict(callbacks=[RichProgressBar()]))"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "75dc8c67",
+   "metadata": {},
+   "outputs": [],
+   "source": []
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "Python 3 (ipykernel)",
+   "language": "python",
+   "name": "python3"
+  },
+  "language_info": {
+   "codemirror_mode": {
+    "name": "ipython",
+    "version": 3
+   },
+   "file_extension": ".py",
+   "mimetype": "text/x-python",
+   "name": "python",
+   "nbconvert_exporter": "python",
+   "pygments_lexer": "ipython3",
+   "version": "3.9.16"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 5
+}
diff --git a/examples/pangeo-tem.ipynb b/examples/pangeo-tem.ipynb
deleted file mode 100644
index 222b2f4d..00000000
--- a/examples/pangeo-tem.ipynb
+++ /dev/null
@@ -1,2711 +0,0 @@
-{
- "cells": [
-  {
-   "cell_type": "markdown",
-   "id": "67c8663d",
-   "metadata": {},
-   "source": [
-    "# Pangeo TEM Workload\n",
-    "\n",
-    "This is a notebook for exploring a simplified version of the example in https://github.com/dcherian/ncar-challenge-suite/blob/main/tem.ipynb.\n",
-    "\n",
-    "See https://github.com/tomwhite/cubed/issues/145"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 1,
-   "id": "0f1e3c3c",
-   "metadata": {},
-   "outputs": [],
-   "source": [
-    "import cubed\n",
-    "import cubed.array_api as xp\n",
-    "import cubed.random\n",
-    "\n",
-    "import pandas as pd\n",
-    "import xarray as xr"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 2,
-   "id": "8e1f099a",
-   "metadata": {},
-   "outputs": [],
-   "source": [
-    "dims = (\"time\", \"level\", \"lat\", \"lon\")\n",
-    "# 1 is number of years, adjust to make bigger, \n",
-    "# full dataset is 60-ish years.\n",
-    "# shape = (1 * 365 * 24, 37, 72, 1440)\n",
-    "shape = (1 * 31 * 24, 37, 72, 144)  # make smaller while exploring\n",
-    "chunks = (24, 15, -1, -1)\n",
-    "\n",
-    "spec = cubed.Spec(max_mem=1_000_000_000)"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 3,
-   "id": "b4c7ff6a",
-   "metadata": {},
-   "outputs": [
-    {
-     "data": {
-      "text/html": [
-       "
\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "
<xarray.Dataset>\n",
-       "Dimensions:  (time: 744, level: 37, lat: 72, lon: 144)\n",
-       "Coordinates:\n",
-       "  * time     (time) datetime64[ns] 2001-01-01 ... 2001-01-31T23:00:00\n",
-       "Dimensions without coordinates: level, lat, lon\n",
-       "Data variables:\n",
-       "    U        (time, level, lat, lon) float64 ...\n",
-       "    V        (time, level, lat, lon) float64 ...\n",
-       "    W        (time, level, lat, lon) float64 ...\n",
-       "    T        (time, level, lat, lon) float64 ...
"
-      ],
-      "text/plain": [
-       "\n",
-       "Dimensions:  (time: 744, level: 37, lat: 72, lon: 144)\n",
-       "Coordinates:\n",
-       "  * time     (time) datetime64[ns] 2001-01-01 ... 2001-01-31T23:00:00\n",
-       "Dimensions without coordinates: level, lat, lon\n",
-       "Data variables:\n",
-       "    U        (time, level, lat, lon) float64 ...\n",
-       "    V        (time, level, lat, lon) float64 ...\n",
-       "    W        (time, level, lat, lon) float64 ...\n",
-       "    T        (time, level, lat, lon) float64 ..."
-      ]
-     },
-     "execution_count": 3,
-     "metadata": {},
-     "output_type": "execute_result"
-    }
-   ],
-   "source": [
-    "ds = xr.Dataset(\n",
-    "    {\n",
-    "        \"U\": (dims, cubed.random.random(shape, chunks=chunks, spec=spec)),\n",
-    "        \"V\": (dims, cubed.random.random(shape, chunks=chunks, spec=spec)),\n",
-    "        \"W\": (dims, cubed.random.random(shape, chunks=chunks, spec=spec)),\n",
-    "        \"T\": (dims, cubed.random.random(shape, chunks=chunks, spec=spec)),\n",
-    "    },\n",
-    "    coords={\"time\": pd.date_range(\"2001-01-01\", periods=shape[0], freq=\"H\")},\n",
-    ")\n",
-    "ds"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 4,
-   "id": "dd80eff1",
-   "metadata": {},
-   "outputs": [
-    {
-     "data": {
-      "text/html": [
-       "
\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "
<xarray.DataArray 'U' (time: 744, level: 37, lat: 72, lon: 144)>\n",
-       "cubed.Array<array-004, shape=(744, 37, 72, 144), dtype=float64, chunks=((24, 24, 24, 24, 24, 24, 24, 24, 24, 24, 24, 24, 24, 24, 24, 24, 24, 24, 24, 24, 24, 24, 24, 24, 24, 24, 24, 24, 24, 24, 24), (15, 15, 7), (72,), (144,))>\n",
-       "Coordinates:\n",
-       "  * time     (time) datetime64[ns] 2001-01-01 ... 2001-01-31T23:00:00\n",
-       "Dimensions without coordinates: level, lat, lon
"
-      ],
-      "text/plain": [
-       "\n",
-       "cubed.Array\n",
-       "Coordinates:\n",
-       "  * time     (time) datetime64[ns] 2001-01-01 ... 2001-01-31T23:00:00\n",
-       "Dimensions without coordinates: level, lat, lon"
-      ]
-     },
-     "execution_count": 4,
-     "metadata": {},
-     "output_type": "execute_result"
-    }
-   ],
-   "source": [
-    "ds.U"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 5,
-   "id": "f5023814",
-   "metadata": {},
-   "outputs": [
-    {
-     "data": {
-      "text/html": [
-       "
\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "
<xarray.Dataset>\n",
-       "Dimensions:  (time: 744, level: 37, lat: 72)\n",
-       "Coordinates:\n",
-       "  * time     (time) datetime64[ns] 2001-01-01 ... 2001-01-31T23:00:00\n",
-       "Dimensions without coordinates: level, lat\n",
-       "Data variables:\n",
-       "    U        (time, level, lat) float64 ...\n",
-       "    V        (time, level, lat) float64 ...\n",
-       "    W        (time, level, lat) float64 ...\n",
-       "    T        (time, level, lat) float64 ...
"
-      ],
-      "text/plain": [
-       "\n",
-       "Dimensions:  (time: 744, level: 37, lat: 72)\n",
-       "Coordinates:\n",
-       "  * time     (time) datetime64[ns] 2001-01-01 ... 2001-01-31T23:00:00\n",
-       "Dimensions without coordinates: level, lat\n",
-       "Data variables:\n",
-       "    U        (time, level, lat) float64 ...\n",
-       "    V        (time, level, lat) float64 ...\n",
-       "    W        (time, level, lat) float64 ...\n",
-       "    T        (time, level, lat) float64 ..."
-      ]
-     },
-     "execution_count": 5,
-     "metadata": {},
-     "output_type": "execute_result"
-    }
-   ],
-   "source": [
-    "# Note: we actually want skipna=True, but this isn't implemented in xarray yet\n",
-    "# see https://github.com/pydata/xarray/issues/7243\n",
-    "zonal_means = ds.mean(\"lon\", skipna=False)\n",
-    "zonal_means"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 6,
-   "id": "02788a25",
-   "metadata": {},
-   "outputs": [
-    {
-     "data": {
-      "text/html": [
-       "
\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "
<xarray.Dataset>\n",
-       "Dimensions:  (time: 744, level: 37, lat: 72, lon: 144)\n",
-       "Coordinates:\n",
-       "  * time     (time) datetime64[ns] 2001-01-01 ... 2001-01-31T23:00:00\n",
-       "Dimensions without coordinates: level, lat, lon\n",
-       "Data variables:\n",
-       "    U        (time, level, lat, lon) float64 ...\n",
-       "    V        (time, level, lat, lon) float64 ...\n",
-       "    W        (time, level, lat, lon) float64 ...\n",
-       "    T        (time, level, lat, lon) float64 ...\n",
-       "    uv       (time, level, lat, lon) float64 ...\n",
-       "    vt       (time, level, lat, lon) float64 ...\n",
-       "    uw       (time, level, lat, lon) float64 ...
"
-      ],
-      "text/plain": [
-       "\n",
-       "Dimensions:  (time: 744, level: 37, lat: 72, lon: 144)\n",
-       "Coordinates:\n",
-       "  * time     (time) datetime64[ns] 2001-01-01 ... 2001-01-31T23:00:00\n",
-       "Dimensions without coordinates: level, lat, lon\n",
-       "Data variables:\n",
-       "    U        (time, level, lat, lon) float64 ...\n",
-       "    V        (time, level, lat, lon) float64 ...\n",
-       "    W        (time, level, lat, lon) float64 ...\n",
-       "    T        (time, level, lat, lon) float64 ...\n",
-       "    uv       (time, level, lat, lon) float64 ...\n",
-       "    vt       (time, level, lat, lon) float64 ...\n",
-       "    uw       (time, level, lat, lon) float64 ..."
-      ]
-     },
-     "execution_count": 6,
-     "metadata": {},
-     "output_type": "execute_result"
-    }
-   ],
-   "source": [
-    "anomaly = ds - zonal_means\n",
-    "anomaly['uv'] = anomaly.U*anomaly.V\n",
-    "anomaly['vt'] = anomaly.V*anomaly.T\n",
-    "anomaly['uw'] = anomaly.U*anomaly.W\n",
-    "anomaly"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 7,
-   "id": "1716e1c4",
-   "metadata": {},
-   "outputs": [
-    {
-     "data": {
-      "text/html": [
-       "
\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "
<xarray.Dataset>\n",
-       "Dimensions:  (time: 744, level: 37, lat: 72)\n",
-       "Coordinates:\n",
-       "  * time     (time) datetime64[ns] 2001-01-01 ... 2001-01-31T23:00:00\n",
-       "Dimensions without coordinates: level, lat\n",
-       "Data variables:\n",
-       "    U        (time, level, lat) float64 ...\n",
-       "    V        (time, level, lat) float64 ...\n",
-       "    W        (time, level, lat) float64 ...\n",
-       "    T        (time, level, lat) float64 ...\n",
-       "    uv       (time, level, lat) float64 ...\n",
-       "    vt       (time, level, lat) float64 ...\n",
-       "    uw       (time, level, lat) float64 ...
"
-      ],
-      "text/plain": [
-       "\n",
-       "Dimensions:  (time: 744, level: 37, lat: 72)\n",
-       "Coordinates:\n",
-       "  * time     (time) datetime64[ns] 2001-01-01 ... 2001-01-31T23:00:00\n",
-       "Dimensions without coordinates: level, lat\n",
-       "Data variables:\n",
-       "    U        (time, level, lat) float64 ...\n",
-       "    V        (time, level, lat) float64 ...\n",
-       "    W        (time, level, lat) float64 ...\n",
-       "    T        (time, level, lat) float64 ...\n",
-       "    uv       (time, level, lat) float64 ...\n",
-       "    vt       (time, level, lat) float64 ...\n",
-       "    uw       (time, level, lat) float64 ..."
-      ]
-     },
-     "execution_count": 7,
-     "metadata": {},
-     "output_type": "execute_result"
-    }
-   ],
-   "source": [
-    "temdiags = zonal_means.merge(anomaly[['uv','vt','uw']].mean(\"lon\", skipna=False))  # actually want skipna=True\n",
-    "temdiags"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 8,
-   "id": "9250bbfd",
-   "metadata": {},
-   "outputs": [
-    {
-     "data": {
-      "text/html": [
-       "
\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "
<xarray.Dataset>\n",
-       "Dimensions:    (dayofyear: 31, level: 37, lat: 72)\n",
-       "Coordinates:\n",
-       "  * dayofyear  (dayofyear) int64 1 2 3 4 5 6 7 8 9 ... 24 25 26 27 28 29 30 31\n",
-       "Dimensions without coordinates: level, lat\n",
-       "Data variables:\n",
-       "    U          (dayofyear, level, lat) float64 ...\n",
-       "    V          (dayofyear, level, lat) float64 ...\n",
-       "    W          (dayofyear, level, lat) float64 ...\n",
-       "    T          (dayofyear, level, lat) float64 ...\n",
-       "    uv         (dayofyear, level, lat) float64 ...\n",
-       "    vt         (dayofyear, level, lat) float64 ...\n",
-       "    uw         (dayofyear, level, lat) float64 ...
"
-      ],
-      "text/plain": [
-       "\n",
-       "Dimensions:    (dayofyear: 31, level: 37, lat: 72)\n",
-       "Coordinates:\n",
-       "  * dayofyear  (dayofyear) int64 1 2 3 4 5 6 7 8 9 ... 24 25 26 27 28 29 30 31\n",
-       "Dimensions without coordinates: level, lat\n",
-       "Data variables:\n",
-       "    U          (dayofyear, level, lat) float64 ...\n",
-       "    V          (dayofyear, level, lat) float64 ...\n",
-       "    W          (dayofyear, level, lat) float64 ...\n",
-       "    T          (dayofyear, level, lat) float64 ...\n",
-       "    uv         (dayofyear, level, lat) float64 ...\n",
-       "    vt         (dayofyear, level, lat) float64 ...\n",
-       "    uw         (dayofyear, level, lat) float64 ..."
-      ]
-     },
-     "execution_count": 8,
-     "metadata": {},
-     "output_type": "execute_result"
-    }
-   ],
-   "source": [
-    "temdiags_mean = temdiags.groupby('time.dayofyear').mean(skipna=False)\n",
-    "temdiags_mean"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 9,
-   "id": "8ed7be49",
-   "metadata": {},
-   "outputs": [
-    {
-     "data": {
-      "text/plain": [
-       "array([[[ 1.92543754e-03,  8.62170647e-04,  1.67535136e-03, ...,\n",
-       "         -2.50751881e-03,  2.46263741e-03, -4.70656500e-04],\n",
-       "        [-9.69165572e-04, -9.19808903e-04, -8.64318054e-04, ...,\n",
-       "          9.10300914e-04,  7.68390560e-04, -9.60456448e-04],\n",
-       "        [ 6.77793313e-04,  2.03099229e-03,  1.44811866e-03, ...,\n",
-       "          2.01528240e-03, -1.03220836e-03,  1.94811893e-03],\n",
-       "        ...,\n",
-       "        [ 2.14917057e-03,  1.47596143e-04, -7.10794437e-04, ...,\n",
-       "         -2.50238921e-04, -1.38891106e-03,  2.17674787e-03],\n",
-       "        [ 1.19509492e-03, -7.62831655e-04, -1.08836927e-03, ...,\n",
-       "         -1.45508192e-03,  7.55372268e-05,  2.78225531e-04],\n",
-       "        [-4.35496517e-03,  1.28230145e-03,  4.80595941e-04, ...,\n",
-       "          2.88439879e-04,  2.96332222e-03,  7.91863232e-04]],\n",
-       "\n",
-       "       [[ 8.82986835e-04,  1.54489743e-03, -1.49333807e-03, ...,\n",
-       "          5.29418813e-04,  1.63187013e-03, -2.07224193e-04],\n",
-       "        [ 9.55443785e-04,  2.54465546e-03, -9.19900311e-04, ...,\n",
-       "         -1.39288060e-03,  2.21651069e-03,  1.47995069e-04],\n",
-       "        [-7.01814903e-04, -1.68102991e-03, -1.29477818e-03, ...,\n",
-       "          3.13384138e-04, -1.30865629e-03,  1.77824569e-03],\n",
-       "        ...,\n",
-       "        [-3.52397366e-03, -6.79305040e-04, -2.14349293e-03, ...,\n",
-       "          1.87527147e-03, -3.25520479e-05, -2.15899612e-05],\n",
-       "        [ 1.63726090e-03, -5.77533746e-04, -7.36883128e-05, ...,\n",
-       "         -1.39907308e-03, -1.52952078e-04,  8.81376198e-04],\n",
-       "        [-2.35014413e-03, -5.04083084e-04, -4.77540340e-04, ...,\n",
-       "          1.23444797e-03,  5.71524836e-04, -1.06911029e-03]],\n",
-       "\n",
-       "       [[-6.76046164e-04,  1.00822320e-03, -2.24455923e-03, ...,\n",
-       "         -1.10717774e-03, -2.00245065e-03, -7.11825695e-04],\n",
-       "        [ 6.60666716e-04,  8.47521507e-04, -2.09004663e-03, ...,\n",
-       "          1.82431769e-04,  2.98712804e-03,  1.81179336e-04],\n",
-       "        [ 8.90949546e-04,  2.03368887e-04, -8.75401454e-04, ...,\n",
-       "          1.44188907e-03,  7.32184069e-04, -3.16988423e-04],\n",
-       "        ...,\n",
-       "        [ 7.10740482e-04,  1.48801481e-03,  1.15863589e-03, ...,\n",
-       "         -2.00549727e-03, -3.77421384e-05,  9.23212283e-04],\n",
-       "        [ 9.96007282e-04, -1.63505959e-05,  6.54791612e-04, ...,\n",
-       "          6.17042736e-04,  1.29283052e-03, -6.03389263e-04],\n",
-       "        [ 8.24574209e-04,  1.88788382e-03,  3.95372441e-04, ...,\n",
-       "         -9.24382835e-04, -1.27852521e-03, -2.25838836e-05]],\n",
-       "\n",
-       "       ...,\n",
-       "\n",
-       "       [[-1.72225771e-03,  6.66428002e-04, -2.13037730e-03, ...,\n",
-       "         -2.67176501e-03, -4.02366613e-04,  7.35956183e-04],\n",
-       "        [ 3.39723093e-04, -1.29779524e-03, -1.23399194e-03, ...,\n",
-       "         -1.78263750e-03,  1.69794295e-03, -9.93540772e-05],\n",
-       "        [-1.28624363e-03,  3.66594558e-04, -1.50847381e-03, ...,\n",
-       "         -5.46152652e-05, -8.04215342e-04, -7.24984607e-04],\n",
-       "        ...,\n",
-       "        [ 4.10706264e-04, -4.17386006e-04,  2.31336418e-04, ...,\n",
-       "          6.55327474e-04,  4.14411516e-04,  3.24550774e-03],\n",
-       "        [ 1.95525139e-03, -2.35227380e-03, -9.99084542e-04, ...,\n",
-       "          1.62829132e-03,  9.69095656e-04, -1.49439057e-03],\n",
-       "        [ 1.82066333e-03,  6.53274368e-04, -1.86011761e-03, ...,\n",
-       "          1.25957309e-03, -6.88163769e-04,  3.47627300e-04]],\n",
-       "\n",
-       "       [[-2.32699809e-04, -2.00950469e-03,  1.18200895e-03, ...,\n",
-       "          2.91200625e-03, -1.35686054e-03, -1.54041751e-03],\n",
-       "        [-4.38585671e-04, -1.71412560e-04,  1.43469667e-03, ...,\n",
-       "         -8.00626033e-05,  7.56622863e-04,  1.20081303e-03],\n",
-       "        [-1.53841374e-03, -2.13838395e-04, -2.68519284e-03, ...,\n",
-       "          4.40948980e-05,  1.47521752e-03,  2.74396257e-04],\n",
-       "        ...,\n",
-       "        [-6.75613724e-04,  7.10640075e-05, -8.35952594e-04, ...,\n",
-       "         -9.02927942e-04, -2.46258759e-03, -9.94494523e-04],\n",
-       "        [ 2.02874360e-03, -1.89183963e-03,  5.88172912e-05, ...,\n",
-       "         -1.08450790e-03, -7.61052577e-04,  1.69481076e-03],\n",
-       "        [ 3.03819565e-04,  1.49770731e-03,  1.46578274e-03, ...,\n",
-       "          1.05137475e-03,  2.67330224e-03, -1.74462858e-03]],\n",
-       "\n",
-       "       [[-1.09142934e-03,  2.22897568e-03,  1.65366229e-03, ...,\n",
-       "          5.56856096e-04,  1.43595299e-03,  3.46458755e-04],\n",
-       "        [-1.87899240e-03, -9.73849683e-04, -1.16765643e-03, ...,\n",
-       "          1.14138265e-03,  2.20218452e-04, -3.09016418e-06],\n",
-       "        [ 2.04292351e-03, -5.14150022e-04,  7.48065866e-05, ...,\n",
-       "          1.31778954e-03,  1.27937819e-03, -1.78836143e-03],\n",
-       "        ...,\n",
-       "        [ 4.24524791e-04, -2.19092597e-03,  1.05686194e-03, ...,\n",
-       "          1.92356567e-03,  1.24714194e-04, -1.83935611e-03],\n",
-       "        [ 1.35446115e-03, -1.86572888e-05,  8.90174177e-05, ...,\n",
-       "         -7.56328452e-04,  7.03265920e-04, -7.90998972e-04],\n",
-       "        [-1.30668621e-03, -3.58709712e-04,  8.04285067e-04, ...,\n",
-       "         -9.88911246e-04, -4.51279163e-04, -9.39181905e-04]]])"
-      ]
-     },
-     "execution_count": 9,
-     "metadata": {},
-     "output_type": "execute_result"
-    }
-   ],
-   "source": [
-    "temdiags_mean[\"uv\"].values"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": null,
-   "id": "6e152ba3",
-   "metadata": {},
-   "outputs": [],
-   "source": []
-  }
- ],
- "metadata": {
-  "kernelspec": {
-   "display_name": "Python 3 (ipykernel)",
-   "language": "python",
-   "name": "python3"
-  },
-  "language_info": {
-   "codemirror_mode": {
-    "name": "ipython",
-    "version": 3
-   },
-   "file_extension": ".py",
-   "mimetype": "text/x-python",
-   "name": "python",
-   "nbconvert_exporter": "python",
-   "pygments_lexer": "ipython3",
-   "version": "3.8.13"
-  }
- },
- "nbformat": 4,
- "nbformat_minor": 5
-}
diff --git a/examples/pangeo-vorticity.ipynb b/examples/pangeo-vorticity.ipynb
deleted file mode 100644
index bd776418..00000000
--- a/examples/pangeo-vorticity.ipynb
+++ /dev/null
@@ -1,1318 +0,0 @@
-{
- "cells": [
-  {
-   "cell_type": "markdown",
-   "id": "f5a954d5",
-   "metadata": {},
-   "source": [
-    "# Pangeo Vorticity Workload\n",
-    "\n",
-    "This is a notebook for exploring a simplified version of the example in https://github.com/pangeo-data/distributed-array-examples/issues/1.\n",
-    "\n",
-    "In particular, it is the code from https://github.com/coiled/coiled-runtime/issues/174"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 1,
-   "id": "8d745a7b",
-   "metadata": {},
-   "outputs": [],
-   "source": [
-    "import cubed.array_api as xp\n",
-    "import cubed.random\n",
-    "from cubed.extensions.tqdm import TqdmProgressBar"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 2,
-   "id": "1bff01e3",
-   "metadata": {},
-   "outputs": [],
-   "source": [
-    "a = cubed.random.random((1000, 900, 800), chunks=100)\n",
-    "b = cubed.random.random((1000, 900, 800), chunks=100)\n",
-    "x = cubed.random.random((900, 800), chunks=100)\n",
-    "y = cubed.random.random((900, 800), chunks=100)\n",
-    "\n",
-    "result = a[1:] * x + b[1:] * y\n",
-    "result = xp.mean(result)"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 3,
-   "id": "f8ca9389",
-   "metadata": {},
-   "outputs": [
-    {
-     "data": {
-      "image/svg+xml": [
-       "\n",
-       "\n",
-       "\n",
-       "num tasks: 5192\n",
-       "max required memory: 100.0 MB\n",
-       "\n",
-       "\n",
-       "array-017\n",
-       "\n",
-       "\n",
-       "array-017\n",
-       "__getitem__ \n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "array-020\n",
-       "\n",
-       "\n",
-       "array-020\n",
-       "__getitem__ (bw)\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "array-017->array-020\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "array-018\n",
-       "\n",
-       "\n",
-       "array-018\n",
-       "__getitem__ \n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "array-019\n",
-       "\n",
-       "\n",
-       "array-019\n",
-       "__getitem__ (rc)\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "array-018->array-019\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "array-019->array-020\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "array-001\n",
-       "\n",
-       "\n",
-       "array-001\n",
-       "random \n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "array-004\n",
-       "\n",
-       "\n",
-       "array-004\n",
-       "random (bw)\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "array-001->array-004\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "array-002\n",
-       "\n",
-       "\n",
-       "array-002\n",
-       "random \n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "array-003\n",
-       "\n",
-       "\n",
-       "array-003\n",
-       "random (rc)\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "array-002->array-003\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "array-003->array-004\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "array-004->array-020\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "array-021\n",
-       "\n",
-       "\n",
-       "array-021\n",
-       "__mul__ (bw)\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "array-020->array-021\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "array-009\n",
-       "\n",
-       "\n",
-       "array-009\n",
-       "random \n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "array-012\n",
-       "\n",
-       "\n",
-       "array-012\n",
-       "random (bw)\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "array-009->array-012\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "array-010\n",
-       "\n",
-       "\n",
-       "array-010\n",
-       "random \n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "array-011\n",
-       "\n",
-       "\n",
-       "array-011\n",
-       "random (rc)\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "array-010->array-011\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "array-011->array-012\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "array-012->array-021\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "array-028\n",
-       "\n",
-       "\n",
-       "array-028\n",
-       "mean (bw)\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "array-021->array-028\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "array-022\n",
-       "\n",
-       "\n",
-       "array-022\n",
-       "__getitem__ \n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "array-025\n",
-       "\n",
-       "\n",
-       "array-025\n",
-       "__getitem__ (bw)\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "array-022->array-025\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "array-023\n",
-       "\n",
-       "\n",
-       "array-023\n",
-       "__getitem__ \n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "array-024\n",
-       "\n",
-       "\n",
-       "array-024\n",
-       "__getitem__ (rc)\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "array-023->array-024\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "array-024->array-025\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "array-005\n",
-       "\n",
-       "\n",
-       "array-005\n",
-       "random \n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "array-008\n",
-       "\n",
-       "\n",
-       "array-008\n",
-       "random (bw)\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "array-005->array-008\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "array-006\n",
-       "\n",
-       "\n",
-       "array-006\n",
-       "random \n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "array-007\n",
-       "\n",
-       "\n",
-       "array-007\n",
-       "random (rc)\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "array-006->array-007\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "array-007->array-008\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "array-008->array-025\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "array-026\n",
-       "\n",
-       "\n",
-       "array-026\n",
-       "__mul__ (bw)\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "array-025->array-026\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "array-013\n",
-       "\n",
-       "\n",
-       "array-013\n",
-       "random \n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "array-016\n",
-       "\n",
-       "\n",
-       "array-016\n",
-       "random (bw)\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "array-013->array-016\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "array-014\n",
-       "\n",
-       "\n",
-       "array-014\n",
-       "random \n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "array-015\n",
-       "\n",
-       "\n",
-       "array-015\n",
-       "random (rc)\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "array-014->array-015\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "array-015->array-016\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "array-016->array-026\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "array-026->array-028\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "array-029\n",
-       "\n",
-       "\n",
-       "array-029\n",
-       "mean (rc)\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "array-028->array-029\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "array-032\n",
-       "\n",
-       "\n",
-       "array-032\n",
-       "mean (bw)\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       "array-029->array-032\n",
-       "\n",
-       "\n",
-       "\n",
-       "\n",
-       ""
-      ],
-      "text/plain": [
-       ""
-      ]
-     },
-     "execution_count": 3,
-     "metadata": {},
-     "output_type": "execute_result"
-    }
-   ],
-   "source": [
-    "result.visualize(\"pangeo-vorticity\")"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 4,
-   "id": "0b9b8905",
-   "metadata": {},
-   "outputs": [
-    {
-     "data": {
-      "application/json": {
-       "ascii": false,
-       "bar_format": null,
-       "colour": null,
-       "elapsed": 0.00949406623840332,
-       "initial": 0,
-       "n": 0,
-       "ncols": null,
-       "nrows": 47,
-       "postfix": null,
-       "prefix": "array-019",
-       "rate": null,
-       "total": 1,
-       "unit": "it",
-       "unit_divisor": 1000,
-       "unit_scale": false
-      },
-      "application/vnd.jupyter.widget-view+json": {
-       "model_id": "8f268522a58240a180abd2d8f534e021",
-       "version_major": 2,
-       "version_minor": 0
-      },
-      "text/plain": [
-       "array-019:   0%|          | 0/1 [00:00
-
-
-
-
-
-
-num tasks: 5192
-max required memory: 100.0 MB
-
-
-array-017
-
-
-array-017
-__getitem__ 
-
-
-
-
-
-array-020
-
-
-array-020
-__getitem__ (bw)
-
-
-
-
-
-array-017->array-020
-
-
-
-
-
-array-018
-
-
-array-018
-__getitem__ 
-
-
-
-
-
-array-019
-
-
-array-019
-__getitem__ (rc)
-
-
-
-
-
-array-018->array-019
-
-
-
-
-
-array-019->array-020
-
-
-
-
-
-array-001
-
-
-array-001
-random 
-
-
-
-
-
-array-004
-
-
-array-004
-random (bw)
-
-
-
-
-
-array-001->array-004
-
-
-
-
-
-array-002
-
-
-array-002
-random 
-
-
-
-
-
-array-003
-
-
-array-003
-random (rc)
-
-
-
-
-
-array-002->array-003
-
-
-
-
-
-array-003->array-004
-
-
-
-
-
-array-004->array-020
-
-
-
-
-
-array-021
-
-
-array-021
-__mul__ (bw)
-
-
-
-
-
-array-020->array-021
-
-
-
-
-
-array-009
-
-
-array-009
-random 
-
-
-
-
-
-array-012
-
-
-array-012
-random (bw)
-
-
-
-
-
-array-009->array-012
-
-
-
-
-
-array-010
-
-
-array-010
-random 
-
-
-
-
-
-array-011
-
-
-array-011
-random (rc)
-
-
-
-
-
-array-010->array-011
-
-
-
-
-
-array-011->array-012
-
-
-
-
-
-array-012->array-021
-
-
-
-
-
-array-028
-
-
-array-028
-mean (bw)
-
-
-
-
-
-array-021->array-028
-
-
-
-
-
-array-022
-
-
-array-022
-__getitem__ 
-
-
-
-
-
-array-025
-
-
-array-025
-__getitem__ (bw)
-
-
-
-
-
-array-022->array-025
-
-
-
-
-
-array-023
-
-
-array-023
-__getitem__ 
-
-
-
-
-
-array-024
-
-
-array-024
-__getitem__ (rc)
-
-
-
-
-
-array-023->array-024
-
-
-
-
-
-array-024->array-025
-
-
-
-
-
-array-005
-
-
-array-005
-random 
-
-
-
-
-
-array-008
-
-
-array-008
-random (bw)
-
-
-
-
-
-array-005->array-008
-
-
-
-
-
-array-006
-
-
-array-006
-random 
-
-
-
-
-
-array-007
-
-
-array-007
-random (rc)
-
-
-
-
-
-array-006->array-007
-
-
-
-
-
-array-007->array-008
-
-
-
-
-
-array-008->array-025
-
-
-
-
-
-array-026
-
-
-array-026
-__mul__ (bw)
-
-
-
-
-
-array-025->array-026
-
-
-
-
-
-array-013
-
-
-array-013
-random 
-
-
-
-
-
-array-016
-
-
-array-016
-random (bw)
-
-
-
-
-
-array-013->array-016
-
-
-
-
-
-array-014
-
-
-array-014
-random 
-
-
-
-
-
-array-015
-
-
-array-015
-random (rc)
-
-
-
-
-
-array-014->array-015
-
-
-
-
-
-array-015->array-016
-
-
-
-
-
-array-016->array-026
-
-
-
-
-
-array-026->array-028
-
-
-
-
-
-array-029
-
-
-array-029
-mean (rc)
-
-
-
-
-
-array-028->array-029
-
-
-
-
-
-array-032
-
-
-array-032
-mean (bw)
-
-
-
-
-
-array-029->array-032
-
-
-
-
-