Coupling the Annual Atmospheric CO2 Cycle to Satellite NDVI

Published on November 20, 2018 by Jude Kastens

Tags: Animation, Carbon dioxide, China, CO2, NDVI, Remote sensing, United States

China CO2 Terra NDVI

China CO2 Terra NDVI

The annual seasonal cycle in atmospheric CO2 concentrations is well known, largely attributable to the preponderance of mid-latitude, terrestrial vegetation in the northern hemisphere compared to the southern hemisphere. When northern hemisphere vegetation is photosynthetically active, which roughly corresponds with emergence through reproduction and primarily occurs during May-September, atmospheric CO2 is consumed and its concentration declines. Conversely, when northern hemisphere vegetation is senescing or dormant, atmospheric CO2 levels increase. By analogy with organismic respiration, this oscillatory behavior is sometimes described as the Earth “breathing.”

We created two animations illustrating the relationship between CO2 and terrestrial photosynthesis using the famous Mauna Loa CO2 time series synchronized with remotely sensed time-series normalized difference vegetation index (NDVI) data from satellites.

The first animation features China, where the seasonal cycles in vegetative growth patterns are distinct due to the unique and varied geography of the country. The NDVI data used in this video are from the NASA Terra MODIS (Moderate Resolution Imaging Spectroradiometer) 16-day composite dataset – Collection 6. False coloration is applied to the monochromatic NDVI values to provide the impression of “greenness” and help illustrate the vegetation phenology. Also, a static water mask (blue pixels) has been applied to show the location of large water features on the landscape. The time period represented in the video spans approximately March 13, 2000, to September 5, 2018.

China NDVI-CO2 animation

The second animation depicts the conterminous United States, which also harbors pronounced and widespread vegetation seasonality. The NDVI data used in this video are from the older AVHRR (Advanced Very High Resolution Radiometer) sensor installed on several NOAA climatological satellites. Data from multiple AVHRR instruments contributed to this weekly-issued, biweekly composite time series produced by the USGS EROS Data Center. As in the China animation, false greenness coloration is applied to the NDVI values and a static water mask is used. The time period represented in the video spans 1989-2017.

U.S. NDVI-CO2 animation

To be clear, the upward linear trend evident in the CO2 time series is, for the most part, a distinct phenomenon from the seasonal cycles featured in these animations. Also, clouds and snow have not been filtered from the NDVI data. Homogenous patches of light-brown coloration appearing in wintertime likely correspond with snow cover.

The Kansas Biological Survey (KBS) utilizes the U.S. AVHRR NDVI dataset for its operational GreenReport and crop yield forecasting program. Beginning in December 2018, KBS will begin soybean yield forecasting for Brazil and Argentina using the same MODIS NDVI data that are shown in the China animation. For questions regarding the animations, datasets, or the crop yield forecasting program, contact Dr. Jude Kastens.