There is no separate NASA plan. The early Cal/Val period is critical for both of these activities. And with limited resources, we need to work together as a team to get as much out of this cal/val activity as possible.
Miguel Román, PhD NASA Goddard Space Flight Center Pre-Launch Evaluation of the NPP VIIRS Land and Cryosphere EDRs to Meet NASA’s Science Requirements
with contributions from the VIIRS Land Team # : Chris Justice, Ivan Csiszar, Jeff Privette, Mark Friedl, Chengquan Huang, Jerry Zhan, Louis Giglio, Pierre Guillevic, Crystal Schaaf, Dorothy Hall, Tomoaki Miura, Jeff Key, Alexei Lyapustin, Jim Maslanik, George Riggs, Peter Romanov, Igor Appel, Eric Vermote, Wilfrid Schroeder, Robert Wolfe, and Yunyue Yu #Support for the activities underpinning this talk was provided by the JPSS Data Products and Algorithms Division and NASA’s Science Data Segment via the Land PEATE (Product Evaluation and Algorithm Test Element).
The MODIS Land products have had significant uptake by both the science and applications communities.
There is some expectation from these communities that the data stream will continue with VIIRS.
The science team has been providing input to the JPSS program for potential improvements in EDR specifications.
Data Continuity will require Algorithm Continuity
The following is a summary of the VIIRS Land Cal/Val and Science Team’s findings to date (July, 2011) with respect to the utility of the Land and Cryosphere EDRs to meet NASA’s science requirements .
Description: Provides land surface skin temperature at the satellite overpass time under cloud-free conditions.
Multichannel Linear Regression Approach (Atmospheric absorption is corrected using signal differences between the channels in infrared band; a linear regression form is derived from linearization of the radiative transfer process).
Land Surface Temperature (LST) EDR Observed LST is angle dependent Apparent Emissivity Apparent LST
Surface Type Dependency (Coefficients of the regression algorithms are surface type dependent with the IGBP types.)
Current Challenges : LST EDR does not provide dynamic land surface emissivity per the current MODIS “day-night” algorithm. This is a valuable product and should be continued in the JPSS era.
VIIRS LSTs from MODIS as proxy See Bob Yu’s talk: (WE3.T09 – 7/27) LST Retrieval from SEVIRI, as proxy of VIIRS
Product Description : Provides the Normalized Difference Vegetation Index (NDVI) and the Enhanced Vegetation Index (EVI) at 375 m on a daily basis.
Retrieval Strategy :
The NDVI derived from “Top-of-the-Atmosphere (TOA)” reflectance.
The EVI derived from “Top-of-the-Canopy (TOC)” reflectance and based on a 3-band equation.
Current Challenges :
Continuity : The VIIRS NDVI (TOA) will inherently be different from the MODIS NDVI (TOC). The VIIRS EVI equation will be different from the MODIS EVI equation and will require a correction factor for continuity.
Performance : The TOA-NDVI specs were determined for “surface + atmosphere” signals. EVI noise will be influenced by atmospheric corrections (aerosols) and by VIIRS blue band noise.
Algorithm : Temporal compositing will not be part of the IDPS operational algorithm.
TOA-NDVI TOC-EVI MODIS proxy TOA-NDVI & TOC-EVI (2007-177) -0.2 1.0 0 0.5 0.75 0.25 from T. Miura
Description: An ice age classification for the categories: Ice -free, New/Young Ice (less than 30 cm thickness), and All Other ice. Freshwater ice is not included.
An energy budget approach is used to estimate sea ice ice thickness.
Daytime and nighttime algorithms use different approaches.
While the algorithm generates age categories, it actually uses temperature, reflectance and estimated thickness as proxies for age (i.e., age is not calculated directly).
Ice concentration is a intermediate product (IP) in the ice characterization EDR.
Current Challenges : EDR is sensitive to errors in the depth of the snow and the surface albedo. The uncertainty of the product is highest for nighttime retrievals. An alternative model is being investigated.
VIIRS Ice Characterization EDR Examples of ice thickness (left) and ice age (right) over the Arctic. J. Key et al.
Description: Provides Binary Snow Cover and Snow Cover Fraction under cloud-free conditions.
Binary Snow Cover (Uses TOA radiances and the Normalized Difference Snow Index (NDSI) and the VIIRS cloud mask)
Snow Cover Fraction (Uses 2x2 aggregation of the binary snow cover)
Current Challenges : Fractional Snow Cover and Daily Snow Albedo datasets are crucial for monitoring the snowmelt process and surface energy balance at mid- to high latitudes and mountainous areas.
VIIRS Snow Cover EDR NPP_SRFLIIP_L2, surface reflectance Land PEATE , VIIRS PROXY DATA, 2003-033 NPP_VSCM_L2, Binary Snow Cover, Land PEAT, VIIRS PROXY DATA, 2003-033 snow-red, snow-free-green, not snow-blue NPP_VSCD_L2, Snow Cover Fraction, Land PEAT, VIIRS PROXY DATA, 2003-033 snow-red, snow-free-green, not snow-blue from J. Key, P. Romanov, and D. Hall
Description: Geolocation of pixels in which actively burning fires are detected
MODIS Collection 4 hybrid thresholding and contextual algorithm (2003).
On-board aggregation of saturated pixels.
Current Challenges :
VIIRS proxy tests have shown that the ARP produces false fires over arid surfaces.
ARP output lacks the contextual fire mask and fire radiative power (FRP) data layers that are standard components of current and forthcoming active fire data sets (e.g., GOES, and SEVIRI).
The science community addressing global fire emissions, air quality, aerosol studies, and ecosystem processes are currently utilizing (and expect data continuity for) the MODIS Burned Area Product.
Active Fires EDR Examples of NPP VIIRS proxy and corresponding Aqua MODIS active fire data in East-Central Africa on September 6, 2002 AVAFO_npp: 12:25:56 – 12:30:11 MYD14_aqua: 12:25:00 – 12:30:00 VIIRS MODIS L. Giglio (UMCP ) and I. Csiszar (NOAA)
Airborne simulators support prototyping and testing of Level 1 VIIRS Sensor Data Records (SDRs).
- Provide verification of sensor calibration and stability during ICV and LTM.
Aircraft Campaigns in Support of VIIRS Cal/Val Efforts … to Global Land Products (AMS/Ikhana) Simulated Bands: - ETM+ B1-B4 - MODIS B27 - VIIRS M15-M16 (CAR/P3-B) Simulated Bands: - ETM+ B1-B4 - MODIS B1,B2,B3,B5 - VIIRS I2, M3, M5, M7-8 Active Fires + Burned Area Surface Reflectance, VI, BRDF, and Albedo - Development and testing of standard products (L2+). Provides critical in-situ data for multi-sensor validation and intercomparison studies . IFOV: 17.5 mrad FOV: 190.0˚ GIFOV: 4.0-500 m IFOV: 2.5 mrad FOV: 85.9˚ GIFOV: 50 m
If all the Land and Cryosphere EDRs are to serve the needs of the science community, a number of changes to several products and the IDPS algorithm processing chain will be needed.
Other products will also need to be added to the VIIRS Land product suite to provide continuity for all of the MODIS land data record.
With the absence of an AM orbit, it has yet to be determined whether there is sufficient improvement from combining the VIIRS with the METOP AVHRR (1km) , which provides AM data, albeit at 09:30.
As demonstrated by MODIS and the AVHRR, reprocessing of the VIIRS data record will be essential if we are to produce climate quality data records that can further research in Earth System Science.
The capability to produce and distribute a suite of VIIRS science products in formats compatible with MODIS products already exists in the Land PEATE (at-launch processing rate of 12 data-days per day (for two product streams).