To make a classified map remap requires data to train a statistical model called a classifier. Remap uses training points that identify all of the different classes present in the landscape, which it uses to predict across the rest of the landscape. The result is a map where all pixels have been classified into the classes defined by the training set. The classes can be land cover types, such as specific ecosystems (mangrove, forest), forest cover (forest and non-forest), or land uses (agriculture, bare land and urban area).
Any internet connection capable of loading Google Maps will be sufficient to use remap.
Remap is subject to computational limits when running its classification on Google Earth Engine. Therefore, we have set a limit to the focal region of 100,000 km2. If you require maps that are larger than this, we suggest implementing your workflow directly in the Google Earth Engine.
Remap produces classied maps at the Landsat scale (~30m). Individual features on the earth's surface such as trees or small buildings may not be detectable at this scale, so we recommend working with focal region than about one square kilometre.
No. Remap analyses free and open data and the base imagery does not meet these conditions. All classifications in remap use freely available Landsat imagery provided by the USGS. We allow viewing of the high resolution imagery because it can be very useful for developing accurate training sets.
Remap uses the region of interest to contain the classification procedure. It is not possible to do this with more than one polygon.
No. For custom analyses use Google Earth Engine.
If your .csv file with training points is not being recognised by remap, try:
latitude, longitude, label in the same order (
latitude, longitude, label).
Boreal Forest or
Wet Schlerophyll Forest
WGS84, and use
, as the decimal point. A single point should have the format
-34.1524, 151.0259, Wet Schlerophyll Forests.
Remap offers two different ways to download your results in a spatially rectified raster format (.tif):
Direct Download button to download a .zip file with your .tif and a metadata file for your classification. This option can become unreliable with very big downloads, so we recommend using the next method for large classifications.
Drive export geotiff function to download directly to your drive. You will need to first sign into your Google account. Following a period of up to 24 hours, depending on the size of your classification, your results will be delivered to the base folder of your Google Drive. Use this option for big downloads where the direct download approach failed.
The download feature in remap is particularly complex due to the need to obtain results from the Google Earth Engine API. Remap therefore only allows one download to occur at any one time. If your download doesn't work, you can try:
You can save your remap workspace so that you can quickly and easily return to work:
7. Export Data and select
2. Build Training Set, select
Upload Data and
Upload JSON. Remap will zoom to the focal region, import your training set and maintain your colour scheme.
Remap will perform best with a comprehensive and sufficient training set. Try running the classifier with >20 points per class, and if the classification is unsatisfactory simply add more training points..
Yes. Double click on the training point to delete it..
There are a number of choices that can be made to produce the most accurate map classifications. Try these steps to produce the most accurate map possible:
Remap uses a machine learning classification method called Random Forest (RF). RF requires training data (the training set, a set of observations depicting accurate locations of each map class) that are used to teach the classifier the most likely class that a pixel belongs to. Without a sufficient amount of training data, the classifier will perform poorly and may produce a map where pixels appear to be randomly assigned to each class. To overcome this a sufficient training set (>50 point per class) is necessary. Remap will warn you if there are less than 20 training points in any class of your training set.
The classifier predicts class membership of each pixel using the training set. To acheive this, it utilises information from a set of predictor layers. The default set of biophysical (slope, elevation), spectral (Landsat Archive imagery) and climatic (precipitation, temperature) predictors generally perform well across a variety of landscapes. However, you can tune the predictor set to incorporate extra predictors that are expected to strongly influence the distribution of particular land cover types.
Remap will classify the landscape only into the classes you provide it. Therefore, in cases where there are many contrasting map classes, it may be necessary to carefully think through the best set of classes that remap will need to distinguish. For example, in mapping mangrove ecosystems you will get a better results if you map mangroves, water and land (3 classes), rather than simply mangroves and other (2 classes).
Remap was developed to provide access to cutting edge remote sensing methods. However, for some remote sensing problems it remains necessary to develop specialised remote sensing methods. For example, for marine ecosystem types that are inundated by water, or for land cover types that undergo complex temporal changes throughout the growing season, a specialised remote sensing method may be required. In this case consider collaborating with experts in remote sensing or implement your own classification workflow directly in Google Earth Engine.
Remap calculates these results on the fly. In some circumstances the computation may time out due to limits placed on remap by the Google Earth Engine. If the results or assessment tabs do not load, wait for two minutes and try again to access the cached results.
Extent of occurrence (EOO) and area of occupancy (AOO) are two standardized measures of geographic range size used in both IUCN red list criteria. Neither AOO nor EOO is intended to be a precise estimate of total area in which a species or ecosystem occurs. Rather, in risk assessment protocols such as the IUCN red lists they function as standardized, complementary and widely applicable measures of risk-spreading against spatially explicit threats. EOO is measured by a minimum convex polygon encompassing all known occurrences of a species or ecosystem. AOO is an index related to total occupied area, estimated by summing the areas of grid cells of a standard size (~2×2 km for species and 10×10 km for ecosystems. Refer to the IUCN Red List of Ecosystems guidelines or Murray et al (2017) for further information.
Please report the bug to us through our feedback form so we can fix it.
Remap was developed with funding from a Google Earth Engine research award. We appreciate all support feedback to help further develop remap into a tool that helps monitor and conserve the environment.