Canada CFIA - Ottawa

Specific Project Objectives & Deliverables


Two crops were monitored: soybean (F14) and corn (F14N). Soybean and corn were planted by the private producers. April, May and July were drier than normal and June was near normal for rainfall.

Figure 8: Daily precipitation measured at field site (April 12 to October 26) or at the airport site (before April 12 and after October 26).

The delay in the seeding was seen in the biomass sampling data. The rainfall in mid-August resulted in some enhanced growth in the late seeded soybean but not the early seeded soybean.

Figure 9: Dry shoot biomass, plant area index and green leaf area index of soybean in 2016 (CFIA, field 14).

Fraction of absorbed PAR was measured using longbars in early and late seeded soybean as well as corn. PASTiS-PAR and PASTiS-57 measurements were attempted but difficulties with equipment prevented gaining meaningful data.

Figure 10: Fraction of absorbed PAR from longbar measurements.

The delay in seeding was also seen in the leaf chlorophyll measurements obtained with the Dualex.

Figure 11: Dualex leaf chlorophyll readings for the top leaves of soybean. Dualex units approximate µg m-2 leaf chlorophyll.

Soybean yield measurements showed values slightly smaller in the drier areas of the field (west) and slightly higher values in the wetter areas (east, water feature in the centre).

Figure 12: Yield map from soybean harvest, headlands and partial passes removed. Yield expressed as kg m-2


Yield from the corn field has not been finalized. Yield data from the corn and soybean fields have not been received from the private producers.

Eddy covariance measuring systems were recording fluxes during all the growing season. The cumulative evapotranspiration associated with the late seeded soybean (295 mm) was lower than that of the early seeded soybean (317 mm).

The cumulative dry biomass extracted from the CO2 flux data indicated that the late seeded soybean accumulated about 25% less biomass than the early seeded site. 

Biophysical products were derived by CNES from Landsat-8 and RapidEye data acquired over CFIA in 2013 and 2014, using the Sentinel-2 product processing stream.

For Landsat-8 data, top of canopy reflectance in the green, red, NIR and SWIR-1 bands were used; for RapidEye, all five bands (blue, green, red, red-edge and NIR) were used.

The products include the estimates of four biophysical parameters: LAI,  cover fraction (fCover), black and white sky fAPAR.

The accuracy of the products is satisfactory. The two attractive aspects of the S2 products are as follows:

1) the products result from an inversion approach based on radiative transfer modeling;

2) the products issued from the two satellite sensors are consistent, which shows the robustness of the approach knowing that the reflectance information used for the two sensors is quite different.

The performance of assimilation strategies using different optimization algorithms based on JECAM CFIA-Ottawa datasets is in progress.

©2015 Joint Experiment for Crop Assessment and Monitoring