Land Product Validation of MODIS Derived FPAR products over a tropical
dry-forest
Abstract
In remote sensing, being able to ensure the accuracy of the satellite
data being produced remains an issue; this is especially true for
phenological variables such as the Fraction of Photosynthetically Active
Radiation (FPAR). FPAR, which is considered an essential climate
variable by the Global Terrestrial Observation System (GTOS), utilizes
the 400–700 nm wavelength range to quantify the total amount of solar
radiation available for photosynthetic use. It is a variable that is
strongly influenced by the seasonal, diurnal, and optic properties of
vegetation making it an accurate representation of vegetation health.
Measurements of ground level FPAR can be completed using flux towers
along with a limited number of wireless ground sensors, but due to the
finite number and location of these towers, many research initiatives
instead use the Moderate Resolution Imaging Spectroradiometer (MODIS)
FPAR product, which converts Leaf Area Index (LAI) to a FPAR value using
Beer’s Law. This is done despite there being little consensus on whether
this is the best method to use for all ecosystems and vegetation types.
One particular ecosystem that has had limited study to determine the
accuracy of the MODIS derived FPAR products are the Tropical Dry Forests
(TDFs) of Latin America. This ecosystem undergoes drastic seasonal
changes from leaf off during the dry season to green-up during the wet
seasons. This study aims to test the congruency between the MODIS
derived FPAR values and ground-based FPAR values in relation to growing
season length, growing season start and end dates, the peak and mean of
FPAR values, and overall growth/phenological trends at the Santa Rosa
National Park Environmental Monitoring Super Site (SR-EMSS) in Costa
Rica and FPAR MODIS products. We derive our FPAR from a Wireless Sensor
Network (WSN) consisting of more than 50 nodes measuring transmitted
PAR, temperature, relative humidity, and soil moisture over custom time
intervals ranging from 2-Hz to 15 min since 2013. Our fundamental goal
is to demonstrate how accurate and reflective the MODIS derived FPAR
product is of TDF phenology. This will be the first step taken in
identifying potential problems with the MODIS derived FPAR products over
TDFs in the Americas.