Print This Page

JECAM | Joint Experiment for Crop Assessment and Monitoring

Madagascar - Antsirabé

Project Overview

Goals of the site in terms of:

This work aims at testing the potential of the mission SENTINEL-2 to map croplands in a region of Madagascar characterized by small size fields, fragmented farmland and frequent cloud cover. The overall objective of this research is to provide new products from the future satellite mission, based on existing (SPOT, Landsat-8) or recent (PLEIADES) missions to support early warning systems for food security. This preparatory work is conducted in two steps: (i) mapping of different cropping systems from multisource data (SPOT time series, very high resolution PLEIADES images, DEM, ground data) and data mining methods (Random Forest) and (ii) estimation of agricultural production (phenological transition dates, yield).


Project Reports

2016 Site Progress Report

2015 Site Progress Report

2014 Site Progress Report

Journal Articles:

Vintrou, E., Lebourgeois, V., Bégué, A., Ienco, D., Teisseire, M., Todoroff, P., Ramahandry, F., 2014. Crop mapping in complex landscape by multi-source data mining and remote sensing for food security. Sentinel-2 for Science Workshop, Frascati, Italy, 20-22 May (submitted).

Bontemps S., Arias M., Cara C., Dedieu G., Guzzonato E., Hagolle O., Inglada J., Matton N., Morin D., Popescu R., Rabaute T., Savinaud M., Sepulcre G., Valero S., Ahmad L., Bégué A., Bingfang W., de Abelleyra D., Diarra A., Dupuy S., French A., ul Hassan Akhtar I., Kussul N., Lebourgeois V., Le Page M., Newby T., Savin I., Verón S.R., Koetz B., Defourny P. 2015. Building a data set over 12 globally distributed sites to support the development of agriculture monitoring applications with sentinel-2. Remote Sensing, 7 (12) : p. 16062-16090.


Implementation Plans

Data Collected for Crop Characterization

Field surveys were conducted in the study zone during the growing peak (end of February) of the 2014-2015 cropping seasons in order to characterize the main cropping systems. A total of 1020 GPS waypoints were registered in the study area, chosen according to their accessibility and to be as well representative of the existing cropping systems as possible. The data gathered during the field surveys concerned farmers’ practices (type of crop, use of fertilizers and irrigation). GPS waypoints were also registered on different types of non-cropped classes (natural vegetation, urban areas, water bodies…) to obtain data on the non-crop class. 199 additionally points were added to the field database for non-crop class by photo interpretation of PLEAIDES very high resolution imagery, making a total of 1219 points finally usable (860 cropped and 359 non-cropped). The boundaries of the fields (for cropped classes) / objects (for non-cropped classes) were then digitized over PLEAIDES very high resolution imagery in order to obtain a polygon database.

Table 1 presents the number of polygons per type of class for the more detailed level (Sub Class) of the JECAM nomenclature. Green is cropped, and grey is non cropped. The map of the GPS waypoints acquired over the Antsirabe JECAM Site over 2014-2015 growing season is presented in Figure 1.

Table 1   Number of Polygons per Type of Class in the Field Database for Sub-class of JECAM Nomenclature

 

Figure 1   Map of the GPS Waypoints Acquired over the Antsirabe JECAM Site over 2014-2015 Growing Season

Data Collected for Estimate of Crop Production

During 2015 harvest season, 124 rice fields (88 irrigated, 36 rainfed) were sampled in order to obtain information on the farmers’ practices (cultivar, planting date, irrigation, fertilization, harvest date) and the yield (total biomass, dry biomass, full and empty grain yield on two plots of 1m² inside the field). The map in Figure 2 presents the location of the 2015 yield surveys.

Figure 2   Locations of 2015 Field Surveys

Plans for Next Growing Season:

The same approach will be maintained for the next growing season but with time series of Sentinel-2 images (preprocessed by the CNES MUSCATE processing chain to obtain images in top of canopy reflectance and monthly synthesis free from clouds) and SPOT 6/7 images for the very high spatial resolution of the site during the peak of the growing season.

We anticipate ordering the same type/quantity of EO data next year. 


Site Description

Locations

Madagascar - Antsirabé
Site Extent   Centroid: -19.410, 47.087
Top left: -19.384, 47.080 Bottom Right: -20.008, 45.371

The study site (around 60 km * 60 km) is located near Antsirabe, in the Highlands Region of Madagascar. In this region, like in the whole country, the main crop is rice, mainly cultivated under irrigation, on terraces or basins. Nevertheless, rainfed rice started to develop on the hills since the recent spreading of new resistant varieties. Information is missing on cropped surfaces, their distribution and production in a country touched by food insecurity.

Location

Topography

Soils

Drainage class/irrigation

Crop calendar:

Field size

Climate and weather

Agricultural methods used

Photo 1: Irrigated rice Fields on a basin

Photo 2: Rainfed rice and associated maize

Photo 3: Irrigated rice cultivated on terraces

Photo 4: Irrigated Rice(bottom) and Rain Fed Corn (top)

Photo 4: Rice Fields in an Irrigated basin


Specific Project Objectives & Deliverables

Results

The methodology for crop identification and crop area estimation is based on the combined use of object based image analysis and data mining. It involves 3 steps:

  1. Preprocessing steps: (i) All images were transformed to top of atmosphere reflectance (top of canopy reflectance was tested but without satisfactory results because of the absence of reliable information on atmospheric composition), (ii) the boundaries of each field/object of the field database were digitized based on very high resolution PLEAIDES imagery, (iii) SPOT DEM was processed to extract slopes.
  2. Building a learning database:  a learning database was built by extracting a set of 341 variables for each field/object of the field database including: 286 radiometric variables (spectral response, indices), 50 textural variables (only from PLEIADES imagery), and some topographic (altitude, slope) and geometric (object size) variables.
  3. Random Forest: (i) a classifier was built based on the learning database, for each level of the JECAM nomenclature, (ii) an optimization phase was performed by analyzing the importance (informative degree) of each variable used and reducing the amount of variables used to perform the classifications to an optimal volume (providing the best classification performances for each level), (iii) the classifications were performed for the whole site at each level.

 

Following figures present the class and overall accuracies obtained for each level of the JECAM nomenclature, and the maps obtained for the whole area at each level.

 

Table 1   Class and Overall Accuracies Obtained for each Level of the JECAM Nomenclature using Random Forest over the Learning Database

Figure 1   Crop-Non Crop Level

Figure 2   Land Cover Level

Figure 3   Crop Group Level

Figure 4   Sub-Class Level

Estimation of rice crop production

Work on estimation of rice crop production is in progress. Due to the small size of cultivated fields, compared to the spatial resolution of satellite images (10 – 20 m), temporal signal (NDVI from TOA reflectance in Red and PIR bands) extracted for each sampled plot was analyzed and smoothed using Stavitsky-Golay algorithm to eliminate noise linked to mixed pixels, but also to clouds, different sensors, and atmospheric effects. This allowed isolating only plots having a pure temporal signal (by comparing raw temporal signal and smoothed one). A set of twelve satellite variables was then extracted from the NDVI temporal profile of each plot: maximum NDVI of the growing season and integrals of the NDVI on different periods of the growing cycle. These satellite variables were compared to total biomass, dry biomass, full and empty grain yield to analyze the correlations. Results showed that the more the sorting was drastic (elimination of plots having a temporal NDVI profile with too much noise), the more the correlations between satellite and yield variables were good (but with less population). With a sorting leading to a resulting population of 14 irrigated rice plots (over the 88 plots initially available), the best correlations were obtained with the use of the integral of NDVI from the middle of the growing slope to the maximum of NDVI of the growing cycle (here referred to as integral.mid.max). Good and very significant (p ≤ 0.01) linear correlations were obtained between this integral.mid.max and total biomass (R² = 0.68) and straw biomass (R² = 0.61). For grain yield, the correlation was less important (R² = 0.58) but significant (p ≤ 0.01). Empty grain yield showed non-significant correlations with all satellite variables.

The estimation of rice crop production still needs to be analyzed further.


In Situ Observations

  1. Parameter: GPS waypoints
    Data Collection Protocol:

    A total of 1125 GPS waypoints (851 cropped and 274 non-cropped) were registered in the study area, chosen according to their accessibility and to be as well representative of the existing cropping systems as possible (2013-2014)

    Frequency:
  2. Parameter: Bio-mass yield
    Data Collection Protocol:

    Measurements on a one (1) meter square plot in each field

    Frequency: Annual (just before harvest)
  3. Parameter: Nitrogen status(starting 2014)
    Data Collection Protocol:

    Measurements with a chlorophyll-meter(SPAD, Konica Minolta

    Frequency: 6 times(one/month during growing season)
  4. Parameter: Crop type. Planting date, Harvest date, Location in tops squence
    Data Collection Protocol:

    In situ observation

    Frequency: Annual
  5. Parameter: type of crop, use of fertilizers and irrigation
    Data Collection Protocol:

    Fields surveys were conducted in the study zone during the growing peak (end of February) of the 2013-2014 cropping seasons, field surveys concerned farmers’practices

    Frequency:

EO Data Requirements

Approximate Start Date of Acquisition: October
Approximate End Date of Acquisition: June
Spatial Resolution: 0.5m - 20m
Temporal Frequency: 10-15 days
Latency of Data Delivery:
Wavelengths Required: R, NIR
Across Swath: 60km
Along Track: 60km

SAR Data Requirements

Approximate Start Date of Acquisition: N/A
Approximate End Date of Acquisition: N/A
Spatial Resolution: N/A
Temporal Frequency: N/A
Latency of Data Delivery: N/A
Wavelengths Required: N/A
Polarization N/A
Incidence Angle Restrictions: N/A
Across Track: N/A
Along Track: N/A

Locations

Madagascar - Antsirabé

Centroid
Latitude: -19.410
Longitude: 47.087

Site Extent
Top left
Latitude: -19.384
Longitude: 47.087
Bottom Right
Latitude: -20.008
Longitude: 45.371


Optical Sensors

PLEAIDES
Imaging Mode: Bundel (50cm Pan + 2m 4-Band Colour)
Spatial Resolution: 0.5m Pan, 2m multispectral
Acquisition Frequency: annual
Pre-Processing Level: Ortho
Application:

Worldview 2
Imaging Mode: Multispectral
Spatial Resolution: 2m
Acquisition Frequency: 10 days
Pre-Processing Level: Ortho
Application:

SPOT
Imaging Mode: -31 & +31
Spatial Resolution: 10-20m
Acquisition Frequency: 26 images (Oct-Jun)
Pre-Processing Level: Delivered in level 1A then manually orthorectified and converted in top of atmosphere reflectance
Application:

Landsat 8
Imaging Mode: Multispectral
Spatial Resolution: 15m pansharpenned
Acquisition Frequency: 11 times
Pre-Processing Level: Orthorectified and converted in top of atmosphere reflectance manually
Application:

DEIMOS
Imaging Mode: Multispectral
Spatial Resolution: 20m
Acquisition Frequency: 3 times, Nov-Jan 2014
Pre-Processing Level: Orthorectified and converted in top of atmosphere reflectance manually
Application:

Sentinel 2 (2014 Launch date)
Imaging Mode: Multispectral
Spatial Resolution: 10-20m
Acquisition Frequency: 5 days
Pre-Processing Level: Ortho
Application:

JECAM | Joint Experiment for Crop Assessment and Monitoring | Group on Earth Observation

©2013 Joint Experiment for Crop Assessment and Monitoring © HER MAJESTY THE QUEEN IN RIGHT OF CANADA SA MAJESTE LA REINE DU CHEF DU CANADA (2012)