data_specs/IO_LULC_specs.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "d3e362b0-dbd2-424b-93fc-d8c6940b5ae4",
   "metadata": {},
   "source": [
    "# ESRI Global Land Use Land Cover "
   ]
  },
  {
   "cell_type": "raw",
   "id": "e0c1ae98-4f5f-45ba-977d-9c513390348b",
   "metadata": {
    "raw_mimetype": "text/restructuredtext",
    "tags": []
   },
   "source": [
    ".. contents::\n",
    "   :local:"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "0a7cc6df-948d-4559-92f4-39d3f0b6aebc",
   "metadata": {},
   "source": [
    "**Date modified:** 02 August 2022"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c09eee97-0c8a-4f17-93c7-f2bccd6cd48c",
   "metadata": {},
   "source": [
    "## Product Overview"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "3ca72a64-3677-4c0e-a407-fcd07f39174e",
   "metadata": {},
   "source": [
    "### Background"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c03b3437-c6a7-4081-8daf-e49ba3578639",
   "metadata": {},
   "source": [
    "The ESRI global land use land cover (LULC) map at 10 meter resolution was developed by the [Impact Observatory (IO)](https://www.impactobservatory.com/global_maps) with the [Environmental Systems Research Institute (ESRI)](https://livingatlas.arcgis.com/landcover/) and in patnership with [Microsoft AI for Earth](https://www.microsoft.com/en-us/ai/ai-for-earth). \n",
    "\n",
    "Digital Earth Africa provides free and open access to a copy of the ESRI/IO Land Cover product over Africa.\n",
    "\n",
    "A [Jupyter Notebook](https://github.com/digitalearthafrica/deafrica-sandbox-notebooks/blob/main/Datasets/Landcover_Classification.ipynb) which demonstrates loading and using landcover datasets in the Sandbox is also available."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "3d0d5edd-f9af-41c1-9c46-a2eb300cc713",
   "metadata": {},
   "source": [
    "### Specifications"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "9b069fc5-59d1-41c2-9682-a99019391ab0",
   "metadata": {},
   "source": [
    "#### Spatial and temporal coverage"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "2f1fd227-5b30-4bcb-9609-ee82b2a89dd6",
   "metadata": {},
   "source": [
    "Relevant metadata for the ESRI/IO Land Cover product can be viewed on the DE Africa [Metadata Explorer](https://explorer.digitalearth.africa/products/io_lulc)."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "a9160c67-503f-4830-a96d-1bf793a62a85",
   "metadata": {},
   "source": [
    "**Table 1: ESRI/IO Land Cover product specifications**"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "6c7c0d62-4edc-4105-b584-d39e45360653",
   "metadata": {},
   "source": [
    "|Specification | |\n",
    "|----------|-------------|\n",
    "|Product name|  `io_lulc` |\n",
    "|Cell size - X (degrees) | 0.0000898° (~10m)| \n",
    "|Cell size - Y (degrees) | 0.0000898° (~10m) | \n",
    "|Coordinate reference system | `Product has no default CRS.` | \n",
    "|Temporal resolution | Annual | \n",
    "|Temporal range| 2017 - 2021 |\n",
    "|Update frequency| Annual |"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "62fdbe5b-b3af-4e6d-9866-a0393e4b66df",
   "metadata": {},
   "source": [
    "The specific temporal and geographic extents for the product can be explored as an interactive map on the [DE Africa Metadata Explorer](https://explorer.digitalearth.africa/products/io_lulc). \n",
    "Data is available for the region shaded in blue."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "3679367b-0b1b-435d-bbba-c1848d103a53",
   "metadata": {},
   "source": [
    "**Figure 1: ESRI/IO Land Cover product geographic extent**"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "d8b8526b-1a2f-4805-971c-28cb679dca53",
   "metadata": {},
   "source": [
    "<img src=\"../_static/data_specs/IO_LULC_specs/io_lulc_geographic_extent.png\" alt=\"ESRI LULC Geographic Extent\" width=\"300\" align=\"left\"/>"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "bcf766b3-2367-4a97-8633-2612bd866623",
   "metadata": {},
   "source": [
    "#### Measurements"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c78884a5-32b8-4635-b92b-37e20e88bb2c",
   "metadata": {},
   "source": [
    "**Table 2: ESRI/IO Land Cover product measurements**"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "4e2340d9-885c-47ce-906a-e31348a0961d",
   "metadata": {},
   "source": [
    "| Band ID | Description | Units | Data type | No data$^\\dagger$ |\n",
    "|:------------------------------|:--------------|:--------|:--------|---------:|\n",
    "| data | Land cover classification | 1 | uint8 | 0 |"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "adba0ca2-fd22-4d54-8979-be0b353ecb37",
   "metadata": {},
   "source": [
    "### Processing"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b93ccda6-be17-4f9e-a840-6b7d64d8e536",
   "metadata": {},
   "source": [
    "The ESRI/IO global LULC map is derived from the European Space Agency (ESA) Sentinel-2 imagery. \n",
    "A deep learning AI land classification model was trained using a massive training dataset curated by the National Geographic Society. \n",
    "This dataset contains over 5 billion hand-labelled Sentinel-2 imagery pixels, from 6 bands of surface reflectance data: visible blue, green, red, near infrared, and two shortwave infrared bands. \n",
    "The pixels were sampled from over 20,000 sites distributed across all major biomes of the world. \n",
    "\n",
    "The UNET, deep learning model was applied to multiple ESA Sentinel-2 scenes across a year, accessed via Microsoft’s Planetary Computer and scaled using Microsoft Azure Batch. \n",
    "The images were classified into 9 discrete land use/land cover classes. \n",
    "The LULC predictions were then composited to generate a representative map of the year. \n",
    "\n",
    "The original Esri 2020 Land Cover collection uses 10 classes and an older version of the underlying deep learning model. This new map uses an updated model from the 10-class model and combines Grass(formerly class 3) and Scrub (formerly class 6) into a single Rangeland class (class 11). "
   ]
  },
  {
   "cell_type": "markdown",
   "id": "2c830377-198a-40c6-a7ca-c74d9a1f0531",
   "metadata": {},
   "source": [
    "### Media and example images"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "48e89e84-0cd1-454f-962c-b16530f25c55",
   "metadata": {},
   "source": [
    "**Figure 2: ESRI/IO Land Cover classification for Madagascar in 2020**"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "59000d3c-a17e-4561-9190-badd383ebf73",
   "metadata": {},
   "source": [
    "<img src=\"../_static/data_specs/IO_LULC_specs/io-lulc-madagascar-2020.png\" alt=\"IO LULC Geographic Extent\" width=\"700\" align=\"left\"/>"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "7519d955-16f1-4f5a-908c-d8eb0964bbb0",
   "metadata": {},
   "source": [
    "### References"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "ee1b71a1-222b-4ce7-9da2-aa3d83cbd1e0",
   "metadata": {
    "tags": []
   },
   "source": [
    "K. Karra, C. Kontgis, Z. Statman-Weil, J. C. Mazzariello, M. Mathis and S. P. Brumby, \"Global land use / land cover with Sentinel 2 and deep learning,\" 2021 *IEEE International Geoscience and Remote Sensing Symposium IGARSS*, 2021, pp. 4704-4707, doi: 10.1109/IGARSS47720.2021.9553499.\n",
    "\n",
    "Kontgis, C. (2021, June 24). [Mapping the world in unprecedented detail](https://caitlin-kontgis.medium.com/mapping-the-world-in-unprecedented-detail-7c0513205b90)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "da135198-725c-4ce9-8466-dbb85d1e42d6",
   "metadata": {},
   "source": [
    "### License and Acknowledgements"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "481f8842-54ad-4b55-a806-5b428cf07eca",
   "metadata": {},
   "source": [
    "This data is licensed under a Creative Commons by Attribution (CC BY 4.0) license.\n",
    "\n",
    "Credits: Impact Observatory, Microsoft, and Esri"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "1a9d1ba4-af9c-4287-8346-2df56dd88a8b",
   "metadata": {},
   "source": [
    "## Data Acess"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "2be1d378-0d77-4da6-bd7a-42dcbd7429fe",
   "metadata": {},
   "source": [
    "### OGC Web Services (OWS)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "1b829dc7-a1dd-4c9e-b692-8acb125b7111",
   "metadata": {},
   "source": [
    "The ESRI/IO Land Cover product `io_lulc`  is available through the Digital Earth Africa's OWS. "
   ]
  },
  {
   "cell_type": "markdown",
   "id": "e1dbadfa-634b-4cca-8dc2-8305d3c69ff3",
   "metadata": {},
   "source": [
    "**Table 3: OWS data access details.**"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "bb548dd2-75af-42ac-b4aa-372d23ef3511",
   "metadata": {},
   "source": [
    "|OWS details | |\n",
    "|----------|-------------|\n",
    "|Name | `DE Africa Services` |\n",
    "|Web Map Services (WMS) URL | `https://ows.digitalearth.africa/wms?version=1.3.0` |\n",
    "| Web Coverage Service (WCS) URL | `https://ows.digitalearth.africa/wcs?version=2.1.0`|\n",
    "| Layer name | `io_lulc`|"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "bd8f6cf5-63d6-47bb-8099-9cba792d91ba",
   "metadata": {},
   "source": [
    "Digital Earth Africa OWS details can be found at [https://ows.digitalearth.africa/](https://ows.digitalearth.africa/).\n",
    "\n",
    "For instructions on how to connect to OWS, see [this tutorial](../web_services/index.ipynb)."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "929404c6-eb18-4f7e-989a-982d1e293af4",
   "metadata": {
    "tags": []
   },
   "source": [
    "### Open Data Cube (ODC)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "fb174cbf-e68e-4498-b99d-fd5fbbdd76c8",
   "metadata": {},
   "source": [
    "The ESRI/IO Land Cover product can be accessed through the Digital Earth Africa ODC API, which is available through the [Digital Earth Africa Sandbox](https://sandbox.digitalearth.africa/hub/login).\n",
    "\n",
    "**ODC product name:** `io_lulc`"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "97de7070-4c2e-4f52-ad07-d6d4bfa35bdb",
   "metadata": {},
   "source": [
    "The `io_lulc` product has only one specific band of data which can be called by using the default name, `data`,  or by the band’s alternative name, `classification`, as listed in the table below. \n",
    "ODC `Datacube.load` commands without specified bands will load the  `data`  band."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "cede81bd-370a-41bf-87c4-436f75019dc3",
   "metadata": {},
   "source": [
    "**Table 4: ODC product io_lulc band names.**"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "0056560e-fa48-4666-89da-e7b682e1b1a5",
   "metadata": {},
   "source": [
    "|Band name| Alternative names| Fill value |\n",
    "|:------------------------------|:-------------------|---------:|\n",
    "| data                | classification         |      `0` |"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "6c428a8c-2a61-4b79-a01a-167a64f2ab92",
   "metadata": {},
   "source": [
    "## Technical information"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "700dd66f-1d35-4ef4-9056-379d2d2b4130",
   "metadata": {},
   "source": [
    "### Accuracy assessment of the ESRI 2020 Land Cover\n",
    "\n",
    "The original Esri 2020 Land Cover collection uses 10 classes and an older version of the underlying deep learning model.\n",
    "\n",
    "Following best practices for accuracy assessment, Impact Observatory adjusted the acreage estimates for each land cover class in the ESRI 2020 Land Cover product using its respective user’s accuracy as computed from the comparison to the validation set. \n",
    "The deep learning land cover classification model achieved an overall accuracy of 86% on the validation set.\n",
    "This approach also allowed Impact Observatory to produce a 95% confidence interval for each acreage estimate, providing users with a clearer picture of the accuracy and total area for each class. (From the [ESRI release page](https://www.arcgis.com/home/item.html?id=d6642f8a4f6d4685a24ae2dc0c73d4ac))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "92b71d3f-ba94-4e57-a3f5-a47d6ae9c32b",
   "metadata": {},
   "source": [
    "**Figure 3: Confusion matrix of pixel counts evaluated against \"three expert strict\" gold standard validation tiles**"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "d791352e-0ba8-4b16-8ecf-3a16f38f4d87",
   "metadata": {},
   "source": [
    "<img src=\"../_static/data_specs/IO_LULC_specs/io-lulc-accuracy-assessment-2020.png\" alt=\"IO LULC Geographic Extent\" width=\"1000\" align=\"left\"/>"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "7f6f4fb8-7107-436d-8962-00e24112fdf4",
   "metadata": {},
   "source": [
    "*Image courtesy of ESRI.*"
   ]
  }
 ],
 "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": "3.8.10"
  },
  "widgets": {
   "application/vnd.jupyter.widget-state+json": {
    "state": {},
    "version_major": 2,
    "version_minor": 0
   }
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}