Jan 10, 2003
Some MODIS products and pre-processing
Hongjie Xie
New Mexico Tech
Introduction
The MODIS/Terra data products have two different datasets available for downloading: MODIS L1B (geolocation and radiometric calibrated data, through http://daac.gsfc.nasa.gov/data/dataset/MODIS/01_Level_1/03_Cal_Rad_1km/index.html), and MODIS geophysics products (L2, L3 and L4 through http://edcimswww.cr.usgs.gov/pub/imswelcome/). Here I am introducing the difference of these two datasets and using datasets of June 16, 2002 as an example. I concluded that it is difficult to directly use the MODIS data or products for our ET estimates using SEBAL algorithm even with lots of following pre-processing.
1 MODIS L1B dataset
MODIS L1B dataset is daily data (two images for each day: day time and night time images) with 36 bands (actually 38 images because bands 13 and 14 have both low and high gain images). For our purpose, I only downloaded the day time and 1km resolution bands, which is from band 8 to band 36. Of them, bands 8, 9, 10, 11, 12, 13lo, 13 hi, 14lo, 14hi, 15, 16, 17, 18, 19, and 26 are reflective solar bands (GSUB1.A2002167.1755.003.2002351204005.hdf), while 20- 25 and 27-36 are emissive bands (GSUB1.A2002167.1755.003.2002353195012.hdf). Using the ENVI 3.6 or Erdas IMAGINE 8.6 can import and georeference to a particular projection that users want. Pixel numbers of these images are scaled integers. The equation for transferring the scaled integers into reflectance or radiance is scale *(scaled integer-offset). The bellows are the bands and related scales and offsets (can be obtained by using a free software called HDFLook):
band_names: 8,9,10,11,12,13lo,13hi,14lo,14hi,15,16,17,18,19,26
radiance_scales: 0.00795877 0.00492686 0.00363404 0.00281856 0.00224682 0.00112017 0.000833487 0.00110254 0.000614436 0.000929919 0.000895564 0.00663277 0.00907252 0.00660929 0.00289308
radiance_offsets: 316.972 316.972 316.972 316.972 316.972 316.972 316.972 316.972 316.972 316.972 316.972 316.972 316.972 316.972 316.972
radiance_units: Watts/m^2/micrometer/steradian
reflectance_scales: 1.47881e-05 8.39465e-06 5.95066e-06 4.85232e-06 3.85158e-06 2.34699e-06 1.74632e-06 2.37137e-06 1.32155e-06 2.32964e-06 2.98559e-06 2.30157e-05 3.36832e-05 2.45521e-05 2.57143e-05
reflectance_offsets: 316.972 316.972 316.972 316.972 316.972 316.972 316.972 316.972 316.972 316.972 316.972 316.972 316.972 316.972 316.972
reflectance_units: none
band_names: 20,21,22,23,24,25,27,28,29,30,31,32,33,34,35,36
radiance_scales: 6.2624e-05 0.00314951 6.9216e-05 7.9104e-05 3.15561e-05 5.63982e-05 0.000117557 0.00019245 0.000532487 0.000406323 0.000840022 0.000729698 0.000262264 0.000200696 0.000176708 0.000118339
radiance_offsets: 2730.58 2730.58 2730.58 2730.58 1077.44 1560.33 2730.58 2317.49 2730.58 1560.33 1577.34 1658.22 2501.3 2501.3 2501.3 2501.3
radiance_units: Watts/m^2/micrometer/steradian
Dataset files:
1. g167b8-36refl-radi.img (the image with 31 layers after transferred g167b8-36.img: scaled integers into reflectance or radiance using the above equation and parameters)
2 MODIS geophysical products
MODIS geophysical products derived from L1B daily data, but mostly in 8, 16 or 96 days composite data. For our purpose, I downloaded 6 different series of datasets:
· MODIS/TERRA ALBEDO 16-DAY L3 GLOBAL 1KM ISIN GRID V003 (MOD43B3)
· MODIS/TERRA BRDF/ALBEDO MODEL-1 16-DAY L3 GLOBAL 1 KM ISIN GRID V003 (MOD43B1)
· MODIS/TERRA LAND COVER TYPE 96-DAY L3 GLOBAL 1KM ISIN GRID V003 (MOD12)
· MODIS/TERRA LAND SURFACE TEMPERATURE/EMISSIVITY DAILY L3 GLOBAL 1KM ISIN GRID V003 (MOD11)
· MODIS/TERRA LEAF AREA INDEX/FPAR 8-DAY L4 GLOBAL 1KM ISIN GRID V003 (MOD15)
· MODIS/TERRA SURFACE REFLECTANCE DAILY L2G GLOBAL 500M ISIN GRID V003 (MOD09)
· MODIS/TERRA VEGETATION INDICES 16-DAY L3 GLOBAL 1KM ISIN GRID V003 (MOD13)
Two MODIS granules (0805 and 0905) cover the state of New Mexico. I used the MODIS’s reprojection tool (MRT, free software) to mosaic two granules and reproject it to UTM zone 13 and WGS84 coordinate system. And HDFLook tool was used to obtain the related scale and offset for each file.
2.1 Two types of 16-day albedo: white sky and black sky (local solar noon), are provided. Each includes 10 bands, which are derived in seven spectral bands, as well as in the visible (0.3-0.7), the near/mid-infrared (0.7-5.0) and the total shortwave (0.3-5.0) broadbands. From the scaled integer to albedo using the equation: scaled integer * 0.001.
Datasets:
2. m_16day_wsky_albedo.img (10 layers)
3. m_16day_bsky_albedo.img (10 layers)
2.2 Land cover type 1 is specified into 16 classes as:
water 0
evergreen needleleaf forest 1
evergreen broadleaf forest 2
deciduous needleleaf forest 3
deciduous broadleaf forest 4
mixed forests 5
closed shrubland 6
open shrubland 7
woody savannas 8
savannas 9
grasslands 10
permanent wetlands 11
croplands 12
urban and built-up 13
cropland/natural vegetation mosaic 14
snow and ice 15
barren or sparsely vegetated 16
unclassified 254
Dataset:
4. m_landcover.tif
2.3 Temperature including daily day-time and night-time temperatures. And emissivity is the average of the emissivity band 31 and band 32. From the scaled integer to temperature and emissivity using equations: scaled integer *0.002+0.49 for emissivity, and scaled integer *0.02 for temperature.
Datasets:
5. m_daily_temperture.img (layer 1: day-time, layer 2: night-time)
6. m_daily_emissivity.img (averaging the bands 31 and 32)
2.4 Leaf area index equal to scaled integer * 0.1
Datasets:
7. m_8day_lai.img
2.5 Surface reflectance includes the MODIS bands 1-7, and resample to 1 km for our purpose. The equation is scaled integer * 0.0001.
Datasets:
8. m_daily_reflectance.img
2.6 Vegetation index equals to scaled integer * 0.0001
Datasets:
9. m_16day_ndvi.img