1. Due to publishing restrictions, please submit your email address and we'll send you the PDF that you've requested. We will never sell or abuse your personal information.


Frequently Asked Questions

 


 

I am having trouble installing CLASlite – what can I do?

 

Start by checking that you are using Windows 8, Windows 7, Windows Vista, or Windows XP on a 32-64bit machine.  This is a requirement for CLASlite.  Next, send your problem to the claslite_techsupport@carnegiescience.edu e-mail; we will answer as soon as possible.

 

How does variable terrain affect the CLASlite analysis?

 

In general, there is little effect of varying terrain on CLASlite results unless shadows are cast by mountains or hill slopes onto the forest canopy.  At that point, CLASlite results can be unreliable.  Fortunately, the error images provides in standard CLASlite output provide a means to easily locate, assess and discard areas where mountain shadows have created uncertain results.  However, in varying terrain where the terrain remains generally well illuminated, the Monte Carlo Unmixing (AutoMCU) core module is able to “soak up” or accommodate terrain variability, allowing for a good result.

 

How would riverine forests vs. inland forests be shown in the CLASlite output image, would there be any clearly identifiable difference? 

 

Riverine forests often have higher bare substrate or NPV fractions because river inundation either damages the forest or deposits sediments in these forests.  It is typical to see yellow colors in a bare-PV-NPV composite image from the fractional cover mapping step – yellow indicates mixtures of bare and PV.  In addition, the pattern of forest canopy cover (PV) is usually visually distinct from upload “terra firme” forests: the riverine forests often have a striated appearance due to fluvial dynamics.

 

I don’t have the metadata – what can I do?

 

In some cases, the metadata for an image may be unavailable or incomplete.  A common type of metadata that may be lacking for Landsat-7 imagey is the gain settings.  In most cases for tropical forest, the only two gain settings for Landsat-7 will be with all 6 bands having the high gain setting (i.e., HHHHHH) or with only band 4 having the low setting (e.g., HHHLHH).  The option for different processing systems (NLAPS or LPGS) for Landsat data is not critical, but knowing this can produce a slightly better reflectance image.  Please refer to the webpage http://landsathandbook.gsfc.nasa.gov/handbook/ handbook_htmls/chapter12/htmls/level1_differences.html for more information.  If you cannot find the processing system, we recommend using the default, NLAPS.

 

There are two different radiance scaling options for ALI images.  In case you cannot find the metadata related to this, the best option is also to try each option and look at the reflectance images to see which most accurately represents the spectrum of the vegetation or other feature.

 

How many bits per pixel are supported by CLASlite?

 

The type of data required of the input to CLASlite varies, depending on the satellite.  For example, for Landsat data, the values should be 1 byte per pixel.  Please refer to Appendix II (Data Preparation) of the User Guide for details of how the data for the various types of satellite data should be formatted for CLASlite.

 

How is the atmospheric correction done to CLASlite images?

 

CLASlite ingests raw satellite imagery and applies sensor gains and offsets to derive exo-atmospheric radiance for each image band.  The radiance data are then passed to a fully automated version of the 6S atmospheric radiative transfer model (Vermote et al. 1997) to derive apparent surface reflectance for each spectral band.  The 6S model requires a number of inputs which include an estimate of aerosol and water vapor.  These parameters are held in geographic look-up tables derived from the NASA MODIS 1-degree atmospheric products.  Coordinating the MODIS aerosol and water vapor data with the high resolution satellite being processed by CLASlite is done on an automated basis.  CLASlite uses the latest version of 6S (http://6s.ltdri.org/) which supports Landsat-4, 5, and 7, as well as ASTER, ALI, and SPOT.  CLASlite does not do atmospheric correction on MODIS land imagery because those data are already processed to surface reflectance in the 8-day composite product (MOD09A1).

 

For what years does CLASlite have atmospheric data available for atmospheric correction?  What happens if I process an image from a year for which CLASlite does not have atmospheric data?  Will this affect my results?

 

CLASlite contains MODIS aerosol and water vapor data from 2000-present.  Atmospheric data is updated and made available to CLASlite users monthly.  If you process an image from a year in which CLASlite does not have atmospheric data, CLASlite will automatically select the closest, previous data set for use in atmospheric correction.  Because aerosol and water vapor data are derived from monthly means, atmospheric correction with data from a different year creates a sufficiently accurate reflectance image for application of the Auto-MCU, a robust, probabilistic approach.

 

What file formats does CLASlite support?

 

CLASlite supports files in both ENVI format and GeoTIFF.  Files in either format can be processed by the tool in any step.  Additionally, the user has the option to save any resulting file in either format.  The default file format for CLASlite results is ENVI.  ENVI images consist of two files, one without an extension, the image file, and a second with the extension .hdr, an ancillary information file containing necessary spatial information about the image.  Both files are necessary, but the file without an extension is the file that is loaded into CLASlite or ENVI Freelook.  ENVI files can be opened in both ENVI and ERDAS, while GeoTIFFs are a more universal format.

 

When saving a CLASlite result in GeoTIFF format, it is important that the user defines the output filename without an extension.  Selecting the option, "Save result as GeoTIFF," will add the appropriate extension to the resulting file.  If the user adds an extension to the output filename, CLASlite will not recognize the output file and will show an error at the end of the process.

 

Is it possible to use CLASlite for UTM zone 17S?

 

CLASlite now supports all imagery in the UTM projection.

 

I have an ASTER image that is not working well.  What should I do?

 

It’s possible that the radiance and reflectance images were not created with the the correct gains and offsets. ASTER images have a number of different gain settings that require the user to apply them using his or her own image processing software (ENVI, ERDAS, etc.).  Please refer to Appendix II to see how to find the gain settings of your data and apply them before using the imagery in CLASlite.

 

Would CLASlite explain common errors?

 

Most of the error messages are displayed during their occurrence. In most cases, the user would be brought back to the stage previous to the error occurrence for them to rectify their errors. However, there might be cases wherein error messages would be displayed and on accepting those, CLASlite would shut down. In such cases the user would have to rerun the process again. Some of the common error messages that occur are:

 

"Attempt to subscript CUMU_TOTAL with MAXINDEX is out of range"

 

This error occurs when the user chooses to apply Linear 2% stretching on a piece of chosen area which has more zero valued pixels. The linear 2% stretching depends on the histogram and since most of the values are zero, histogram stretching would not be performed. The users are advised to avoid selecting areas with more zero valued pixels and apply linear 2% stretch on it.

 

 “Number of bands does not match the requirement

 

The different satellite/sensor images used in CLASlite depend on the number of bands in each.  Refer to Appendix II:  Image Preparation and Band Analysis for CLASlite Use for further information.

 

Is it possible to perform the multi image analysis combining images from different satellites?

 

You can perform multi image analysis on images from the same satellite as long as the pixel resolution is the same.

 

What should be the overlap area between images in order to perform a multi-image analysis? When would the non-overlapped area be too large as to make the analysis non-acceptable?

 

In order to perform multi-image analysis on a set of MCU images, there is no specified maximum limit but all the images to be compared must at the least have one pixel in common.

 

When in multi-image mode, what happens if I have variation in the amount of time between images?

 

You can input as many images as you wish.  If the timestep is annual or more frequently, then your deforestation and disturbance maps produced by CLASlite will be closer to gross rates of change.  If your images were collected more than a year apart from one another, then the deforestation and disturbance maps produced by CLASlite will represent longer-term, net rates of change.  It is up to you to interpret the forest changes mapped by CLASlite – you need to consider if you are seeing gross or net changes.  Net forest change can miss forest regrowth, canopy recovery, disturbance, and other processes that may occur between image dates.  Gross forest change resolves all forest gains and losses as they occur.

 

Why does CLASlite not support Landsat MSS?

 

In the past, we have decided not to support Landsat MSS since the instrument is noisy.  However, if there is overwhelming demand for MSS data, we will consider supporting it in an updated version of CLASlite.

 

How much error in the image-to-image registration is allowable by CLASlite?

 

Currently, the multi-image analysis feature in CLASlite requires the user to co-register his/her imagery during the image preparation step to one (1) pixel uncertainty.

 

Does CLASlite v.2 support 64-bit Windows?

 

Yes, but you must install IDL for 32-bit machines first so that CLASlite can run in the 32-bit IDL environment.

 

How did the CLASlite team decide on the decision criteria for forest cover, deforestation and disturbance?

 

These criteria are based on practical experience with CLASlite in humid tropical forest settings in Brazil, Peru, Madagascar, Borneo and Hawaii.  However, we provide the fractional cover values (from Step 3) to the user so that he/she can create decision criteria that best suite his/her needs.  Personalized decision criteria can be implemented using third-party software such as ENVI, ERDAS Imagine, or ArcGIS.

 

Can CLASlite run mosaics of satellite imagery?

 

Yes, you can process a mosaic, but you must have enough computer memory to allow it.  It is generally advisable to run CLASlite on individual images and then mosaic the MCU results.