MAPEO DE LA SUPERFICIE DE AGUA: SÍNTESIS DEL MÉTODO
This section presents a resume of the method developed and applied by MapBiomas Agua. For more information on the methodological details,visit ATBD (Algorithm Theory Base Document)
Presentation
The main objective of MapBiomas Agua is to map the dynamics of surface water throughout the territory of the Amazonian countries, on a monthly and annual basis from 1985 to 2022. The data set is publicly available on a web platform to improve management and use of water resources.
Water surface mapping in the Amazon countries used all Landsat satellite scenes with cloud coverage less than or equal to 70% and a spatial resolution of 30 meters. Mapping was conducted at a subpixel scale (SWSC), with Spectral Mixture Analysis (SMA) and empirical classification rules based on fuzzy logic. The mapping covered the period from 1985 to 2022, on a monthly scale, with a total of 396,000 Landsat scenes processed and analyzed on the Google Earth Engine platform.
Organization and database
The general coordination of MapBiomas Agua is led by Imazon and RAISG while the technical and operational coordination is led by Geokarten. The reconstruction of the monthly historical series of water surface was conducted by specialists from all biomes of the Amazon countries, with the leadership of the following institutions: Fundación Amigos de la Naturaleza FAN (Bolivia), Fundación Gaia Amazonas FGA (Colombia), EcoCiencia (Ecuador), Instituto del Bien Común IBC (Perú), Provita and Wataniba (Venezuela), Alliance of Bioversity International and CIAT (Guayanas and Suriname). The water surface mapping algorithm was developed by Imazon and adapted by MapBiomas Agua in this first stage of work.
The development of the MapBiomas Agua dashboard was led by Geodatin and has relevant contributions from the MapBiomas Agua working group and platform users in the design thinking process.
Three types of products were produced by MapBiomas Agua:
- Monthly and annual water surface maps;
- Water surface transition maps between “Water” and “Non-Water” classes. This product was processed with the annual water surface database;
- Trend maps (increase and decrease) of water surface. This product was calculated from monthly water surface data in 5 km x 5 km grids.
The dashboard is made up of maps, statistics and tools for visualization, analysis and data access. It is possible to view the data on an annual and monthly scale, beyond obtaining it in different territorial units. Finally, the dashboard also has an access link to the MapBiomas Agua data API.
Method
The diagram below illustrates the main steps in classifying surface water and water bodies in the Amazon countries, and comprises a subpixel level surface water classifier (SWSC), decision trees and post-classification procedures to generate annual and monthly surface water data sets.
Figure 1. Classification steps of the water surface and water bodies
Description of classification steps:
- Pre-processing
Consists in the selection of Landsat scenes from the sensors: Landsat 5 Thematic Mapper (TM), Landsat 7 Enhanced Thematic Mapper Plus (ETM+), and Landsat 8 Operational Land Imager (OLI); applying cloud and shadow masking to each scene and excluding scenes with more than 70% cloud cover. The visible, near and mid-infrared spectral bands were selected for the application of the Mixture Spectral Model (MEM). The result of the MEM is a set of compositional bands for each pixel of the Landsat image, for the Vegetation, Non-Photosynthetically Active Vegetation (NPV), Soil, Shade, and Cloud components. Water behaves as a dark body (i.e. low reflectance) in Landsat images and therefore has a high percentage of the Shadow component in the pixel. The edges of lakes, rivers, and humid environments, such as floodplains, present a mixture of Shadow (water), Vegetation, and Soil, which allows the detection of water in environments with these types of materials.
- Water Surface Classification
The original surface water sub-pixel classifier (SWSC) algorithm uses three hi| erarchical binary decision rules (e.g. true, false). Because water absorbs a large part of the electromagnetic radiation, an image with Shade fraction, the combination of GV and Soil and Cloud, is used to classify the pixels as surface water. Additionally, a classification based on independent fuzzy logic (fuzzy rules) is applied, in which the degree of truth/certainty (memberships) that a Landsat pixel is classified as water is determined. The average degree of truth was then calculated to obtain a continuous membership map with values ranging between 0 and 1. Based on these memberships, pixels are classified to produce monthly surface water layers.
By calculating the median pixel memberships among available Landsat scenes for each month, pixels were classified as water based on defined thresholds. Procedures were then applied to restore false negatives and remove false positives, based on temporal metrics. Gap filling was then applied to reclassify as water those pixels that were eventually covered by clouds or within areas where no Landsat scenes existed during a given month, using a combination of two rules: within-year median probability and the decadal median. of the corresponding month. Finally, the presence of cloud shadows or other dark objects in the Landsat scene can also produce false positives in the water classification, so a removal filter was applied to reclassify those pixels as non-water.
Annual surface water maps include an identification between permanent and seasonal water, this classification is based on thresholds corresponding to the number of months in which a pixel is classified as water. For the first case, a frequency >= 6 months is considered, and for the second, a frequency between 1 to 5 months.
Figure 2. Monthly classification process.
