User's Guide

ForWarn is a vegetation change recognition and tracking system that uses moderate resolution satellite (MODIS) data. It provides near real-time change maps for the continental United States (CONUS). Below is a collection of reference information for users of the Forest Change Assessment Viewer (FCAV2). Users of all skill levels can see tips for how to navigate in the Viewer with the linked help videos.

Themes

CONUS Vegetation Monitoring Tools

By default, the “Theme” dropdown will be positioned on “CONUS Vegetation Monitoring Tools.”

The “Theme” dropdown relates to the Map Layers window on the left side of the Viewer. The purpose of the different Theme’s is to control the type of data layers added to the Map Layers table of contents. By default, when the Viewer opens for the first time, the Theme selection is positioned on the “CONUS Vegetation Monitoring Tools”. This Theme adds all the basic map layers to the table of contents, including the different versions of ForWarn’s disturbance detection maps. As part of all the different Theme selections, all basic layers contained in the “CONUS Vegetation Monitoring Tools” are made available to the user in the Map Layers window. Below provides detail of data layers added to the Map Layers window when choosing the different thematic data collections.

This Theme adds the three most recent forest change images to the Map Layers table of contents under the tab “Forest Disturbance Detection Maps (MODIS NRT).” There are six baseline products in all, which are listed under their respective baseline header name and are explained in more detail under Map Layers.

Archived Forest Disturbance Detection Maps (MODIS NRT)

To view forest change images previous to the three most recent, use the “Theme” dropdown and choose “Archived Forest Disturbance Detection Maps (MODIS NRT).”

A new tab will be added to the Map Layers table of contents with the same name (directly underneath “Forest Disturbance Detection Maps (MODIS NRT)”), which groups the six different forest change image baselines going back to 2010.

Phenology-Related Datasets

This "Theme" dropdown also provides access to phenology-related datasets created from the MODIS NDVI historical time series. Visit this page for a description of seasonal and derived parameter products available for display (2010–2013 are currently being updated).

Disturbances Assessed

"Disturbances Assessed" is another theme that may be selected to add layers to the Map Layers table of contents. This theme shows point locations of previously assessed apparent disturbances. A user can view information about those potential disturbances by using the Identify icon . It will display the specific forest change image (using the ShareURL), a description, date, and associated news articles, if available.

Map Layers

Working with Data Layers

By default, the FCAV will open with the most recent one-year baseline forest change product displayed.

To activate or deactivate a data layer, check the box next to it in the Map Layers window. The FCAV will populate the selected data onto the map instantly, and the layer’s legend will appear in the Map Tools window. You can also quickly clear (or deactivate) a data layer by clicking on its legend in the Map Tools window.

More than one layer can be active at a time, but the other active data layers may appear hidden. The gear icon beside the layer name allows you to adjust the transparency of each layer, so that other activated layers are visible. To do this, click the gear and adjust the slider bar to alter the layer transparency. With this function, the relationships between multiple data layers and the underlying geography can be examined.

There are a total of six different forest change products available every eight days that highlight potential forest disturbances, which are under the Forest NDVI Change category. The different forest change products relate to the length of the baseline and the compositing method used to display percent NDVI departure. They are:

Change using a 1-Year Baseline (recent disturbances, 1yr or less);
Early Detect 1-Year Baseline (disturbance resolved in as few as 8-days);
Change using a 3-Year Baseline (3-year old disturbances);
Change using a 5-Year Baseline (5-year old disturbances);
Change using a 10-Year Baseline (10-year old disturbances);
Change using an All-Year Baseline (all disturbances since 2000);
Seasonal Progression Departure (by Hargrove);
The following (2) are located under the "Theme = Seasonal Summaries" drop-down:
- Prominent disturbance "Duration" (produced at two periods during the summer growing season);
- Prominent disturbance "Magnitude" (produced once at the end of the summer growing season);

All change products, except the 'Early Detect', are based on a composited mosaic using maximum NDVI value over a 24-day period. The Early Detect uses an 8-day compositing window, because of the limited number of days from which to acquire the maximum NDVI value, these change images may contain more 'nose' associated with clouds, anomalous artifacts and more apparent NDVI decline.

* VIEW CHANGE PRODUCT DESCRIPTIONS AND BEST USE CASES*

https://drive.google.com/file/d/1hqI014928FMVwo6PjGmxfvx9aombjDY0/view?u...

Types of Data

Below are information and tips for how to use the different types of data available in the FCAV.

Fires

Past and active silvicultural treatments, such as prescribed fire and wildfire, can be observed with layers in the "Fires" tab of the Map Layers window.

Historical Aerial Disturbance Surveys

If there is no indication of drought, a disturbance activity could be insect related. Use the "Historical Aerial Disturbance Surveys (USFS FHM)” tab to explore previously known locations of activity.

Also, try the "EFETAC Pest Proximity" layer that summarizes all previously known ADS locations. The layer is activated by selecting the the Historical Aerial Disturbance Survey layer bar, selecting one of the ADS layers, and then immediately selecting one of the polygons using the “Identify” icon . A summarized ADS report is provided containing: 1)the closest disturbance agents in descending probabilities; 2)the closest agent within 75 km; 3)acres of that disturbance; 4) the tree host; 5)number of disturbance polygons in the 75 km radius; 6)year reported; and 7)probability of identification and/or the most likely agent in decreasing orders of probability.

Notice that the disturbance agent reported by the Pest Proximity layer is linked to Google Trends that reports the number and location of Google searches performed on the specific agent. This type of “crowd-sourcing” can be used to gain insight as to the location and timing of specific forest disturbance agents.

eMODIS cloud Assessment Products

If a rapid or sudden NDVI departure is apparent in a single eight-day image compared to the date of the immediately previous image, it could indicate a departure due to clouds or a sensor anomaly. Check the eModis Cloud Assessment products for the same date(s) as the forest change image(s) and look for signs of lighter-toned reflectance and brighter-toned clouds, or look for sensor patterns in the eMODIS true-color composite.

Does the potential disturbance progress in a naturally growing fashion, growing evenly across image dates? This could indicate the onset of drought, insect activity, or possible anthropogenic causes of NDVI change.

Is the NDVI departure area in grass, shrub, agriculture, snow, or on the margins of these land cover types and uses? This could indicate the onset of drought conditions as witnessed by vegetative response to water stress.

Drought - US Drought Monitor

The U.S. Drought Monitor is jointly produced by the National Drought Mitigation Center at the University of Nebraska-Lincoln, the United States Department of Agriculture, and the National Oceanic and Atmospheric Administration. Map courtesy of NDMC-UNL.

VegDRI Products - Vegetation Drought Response Index

These products are developed by scientists at the U.S. Geological Survey Earth Resources Observation and Science (EROS) Center in collaboration with the National Drought Mitigation Center and the High Plains Regional Climate Center, with sponsorship from the U.S. Department of Agriculture's Risk Management Agency. VegDRI maps are produced every two weeks and combine satellite-based observations of vegetation conditions with climate and environmental.

Other Weather and Climate

TRMM

here

Precipitation Derivatives

Percent of Normal - NOAA PRISM – Parameter elevation Regression on Independent Slopes Model

here

Climate - NOAA PRISM – Parameter elevation Regression on Independent Slopes Model

Physiographic Maps

Under the "Physiographic Maps" layer bar there are ancillary data layers that assist in the vetting of potential disturbance, or recovery dynamics. Visit the websites below for more information.

Agriculture: (USDA) Cropland Data Layer
Hydrography: 7 Day Avg. Streamflow
Decadal Change: Statistically mined from the full set of historical MODIS NDVI datasets since 2000, these provisional research datasets depict areas of deciduous and evergreen, thrive and decline, from all disturbance agents and recovery causes.
Phenological (Eco)Regions

Landcover

Forest Type and Forest Group
Forest Biomass
Carbon Stock
National Land Cover Dataset 2006
2001 GAP Landfire
LANDFIRE Fuel Model 40
LANDFIRE Vegetation

Ecoregions

Bailey’s
Omernik
NEON Domains
FWS LCC

Sometimes more recent aerial imagery is available to help vet land cover change, such as from a state’s newer National Agricultural Imagery Program (NAIP) aerial photography. From Landsat, GloVis is a source to download the most recent Landsat imagery to assist in the interpretation of potential disturbances.

Other sources for potential information:

Map Tools

Masking

When first opening the FCAV, by default the most recent forest change image is displayed showing only forest land cover types.

A Masks tab is located in the Map Tools window at right. Expanding this tab will present a set of masking layers that can be used to isolate specific National Land Cover Database (NLCD)-based land cover types and uses. The masks were created using the latest NLCD dataset, and the majority were resampled to 231 meters to match the MODIS resolution.

Deselecting the "Forest" mask will change the display to present an “all-lands” version of the percent NDVI departure imagery available for CONUS. All NLCD-based land cover and land use types can now be isolated, or used in combinations, to explore percent NDVI change information for any of the forest disturbance detection map products.

Share the Map

The Map Tools window offers the capability to send a URL of a view via your email client to a colleague.

Click the “Share this Map” tab, then highlight and copy the URL (ctrl-c), and paste (ctrl-v) the link into your email program. This will allow your colleague to execute the URL and open the FCAV to the same extent and data layers you were using to explore the specific disturbance event.

Graphing

The "Graph NDVI” tool shows a change in the Normalized Differential Vegetation Index (NDVI) over a 14-year history, with discrete NDVI measurements every eight days.

NDVI values are presented in x and y graph format. NDVI values are plotted along the y axis, and the 11-year time scale, with measurements recorded every eight days, are plotted along the x axis. Changes in NDVI values fluctuate and relate primarily to changes to vegetative condition and vigor, but also can noticeably change due to natural disturbance events and anthropogenic modifications. See “HOW TO INTERPRET ANNUAL NDVI PHENOLOGICAL SIGNATURES” to learn more.

To operate the “Graph NDVI” tool, first zoom to your area of interest, then select the icon displayed in the top left location of the toolbar. Perform a single click of the mouse on your area of interest and a separate window will open with a graph of the NDVI values for that specific location. Try using the “imagery” base map to guide your selection of a point to graph. The “Graph NDVI” feature is available across the entire U.S. for all land cover types and land uses.

In order to provide more detail to the axes and data measurements, you can do the following:

Spread out the Date/Time axis. Simply place the mouse cursor inside the graph window and on the x axis, press and hold SHIFT, and move the cursor left and right to expand and contract the x axis. Release the SHIFT key and “drag” the x axis left and right to see more temporal detail.
Spread out the NDVI axis. Simply place the mouse cursor inside the graph window and on the y axis, press and hold SHIFT, and move the cursor up and down to expand and contract the y axis. Release the SHIFT key and “drag” the y axis up and down to see more detail.
Remember, whether expanding or contracting an axis, be careful to place your cursor on the axis to get the desired effect.

In order to get the optimal large-scale use from the Graphing tool, you can do the following:

Zoom into your area of interest and choose the imagery base map.
Make one of the forest change images active and apply a transparency so you can see through to the aerial photography underneath (see “Working with Data Layers”).
Use the transparent pixels of the forest change image as a guide to select a specific pixels’ land cover composition for which the NDVI history will be returned.
By using the transparent forest change images’ pixels as a template, one can photo-interpret a pixel’s land cover composition before the Graph NDVI tool selection to better understand the character of that pixel’s NDVI history.

How to Interpret Annual Signatures

When interpreting the NDVI signature that the Graphing Tool produces, be aware of the following NDVI-curve response characteristics: the data range and amplitude of NDVI values; the duration of change that NDVI values experience across time; the modality, or period, of recurring NDVI values of similar character; the type of land cover or land use that was selected; the phenological character of increase and/or decrease in NDVI values over time and the date that abrupt changes in NDVI values occur. To illustrate different vegetative land covers, examine a few typical NDVI response curves below:

Pine Plantation (Pinus sp., GA)

Has relatively high average NDVI values
Amplitude is relatively narrow
The annual NDVI value increases and decreases due to photosynthetic response to annual phenology

Coastal Redwoods (Sequoia sempervirens, CA)

Has relatively high and constant NDVI values
Amplitude of change is very narrow
There is slight phenological variation across time

Mixed Hardwood Forest (Adirondack Park, NY)

Has high maximum NDVI values
The data range is relatively high
The NDVI increases and decreases due to annual phenology
The NDVI minimum variability may be due to snow events masking the shrub layer

Prairie Grasslands (National Grasslands, ND)

Has relatively low average NDVI values and a large range
There is a sharp NDVI increase and decrease due to annual phenology
The variability in yearly maximums is due to variations in seasonal precipitation

2006 Derby Wildfire (southern MT)

The pre-fire image is dominated by low density Ponderosa Pine
The post-fire image shows tree killing event led to grass-dominated landscape
Note the greater peakiness and lower winter minimums

Non-Native Invasive Species (Hemlock Wooly Adelgid, eastern TN)

The declining Hemlock composition causes a drop in the winter minimum
The defoliation began during the 2005 growing season
Provides ability to monitor the rate of decline for this area and for new sites

Pine Planting and Growth (central MS)

The mid-level NDVI values are at the beginning of the dataset
The crown closure of the pine plantation reduces the amplitude of the annual phenological signal
Note the trend in increased minimum NDVI values over time and the narrowing of amplitude

Strip-Mining Reclamation (western VA)

The NDVI increases due to the mined area being replanted
The area has not achieved maximum greenness of original forest cover
This tool can provide insights into the ability of reclaimed land to recover its productivity within a specified time frame

Data

WMS

Please visit the Data Access page to learn more about Web Map Services.

WCS

Please visit the Data Access page to learn more about Web Coverage Services. Download this pdf to learn about the download procedures.

Mobile

The Forest Change Assessment Viewer (FCAV) is responsive and mobile-friendly. It can be viewed on various tablet and phone devices, such as iPads and mobile phones.

Frequently Asked Questions

What is the best ForWarn change product to use to see if there are any recent forest disturbances in my area of interest?

The three most recent products are located in the Map Layers box, under the header “Near-Real-Time Change Maps (MODIS NDVI)”. The various products shown reflect conditions compared to different baseline expectations. The 1-year product shows change compared to the same period a year ago, the 5-Year-Max product shows current conditions compared to the highest NDVI of record for that period during any of the last 5 years. Considering baselines over different time scales allows you to distinguish between recent events (by observing how the 1 year change product progresses) and gradual decline that can progress over years. A full description of the NDVI change products can be found here.

What do the colors mean?

Generally, the colors relate to productivity as compared to a previous baseline. Shades of dark blue denote an increase in vegetated condition compared to baseline values. Shades of green to red show where productivity is less. Isolated disturbances such as insect defoliation or fire show up as declines ranging from yellow to red, according to their severity, while recovery in subsequent years shows up as shades of blue that are darker than normal.

More precisely, the colors denote the percent change difference when comparing the current NDVI (Normalized Difference Vegetation Index) image value with the maximum NDVI pixel value over the past eight-, three-, or one-year baseline timelines. NDVI values are calculated via a band ratio of red to infrared bands (or portions of the electromagnetic spectrum) and values range from +1.0 to -1.0.

How is the “Early Detect” product different?

Early Detect uses the same 1 year baseline product used in the 1 year change maps, but it differs in how current NDVI is calculated. Early Detect shows the earliest initial indications of new disturbances up to three weeks before the standard ForWarn products by balancing the best view available with the recentness of satellite observations. This means that change may be observed as recently as a few days after a new product becomes available. However, this approach can result in lower current image quality from clouds and other atmospheric effects. Users are advised to use Early Detect with caution and to compare results with the standard ForWarn products that rely on the best daily satellite observation over the potentially much longer 24 day window.

What is the "Phenotype Max” product?

The "All-Year Phenotype Max" baseline uses a sophisticated approach to predict what the undisturbed values of individual cells should be for a period based on a historical group normal behavior. As group values involve averaging, this baseline effectively buffers inter-annual variability, while still providing a long-term high value that resembles the All Year Max product. Technically speaking, using a non-hierarchical k-means clustering (a statistical grouping technique), each cell was "typed" according to its 46 annual NDVI values. The phenological curve for each year and cell were typed separately, so that pixels fall into one of 5,000 clustered phenotypes each year, 2000–2011, then the year with the maximum area under the annual curve was used as the baseline.

How do I find older products?

Click the "Theme" dropdown (located in the top toolbar of the FCAV), and select “Archived Forest Disturbance Maps.”

A new tab will now be added to the Map Layers window with the same name, and located immediately below the "Forest Disturbance Detection Maps" tab.

The forest disturbance image is covering up the base map and I can’t see where I am viewing.

A settings icon is located beside each data layer in the Map Layers window. Clicking this icon will open a control window, allowing you to use the available slider to apply a transparency setting for any data layer. Set the slider at a position where you can see both the base map and the colors of the forest disturbance image.

How do I save or share a view of a potential disturbance with someone else?

The fastest method is to use the "Share this Map" option, located in the Map Tools Window. You can also use the "Print this Map" icon from the toolbar. Finally, you can issue a keystroke on the “print screen” key on your keyboard.

What should I do if I find a forest disturbance in my area?

The first step is to determine if the cause of NDVI decline is due to cloud contamination or a sensor or image processing anomaly. Start by using the MODIS Cloud Products located in either the default theme data collection or the second collection, the “Archived Forest Disturbance Maps.”

If this isn’t apparent, then check to see if the changed areas are primarily located on the margin of the forest mask, or snow pack margin. Do the pixels of the forest change image contain a high grass component? To determine this, use the imagery base map in the FCAV, or load into your desktop GIS and use your states’ recent NAIP imagery to view the most up-to-date aerial photography, or download a recent Landsat image via GloVis.

If it is still undetermined, use the FCAV’s weather and climate layers to see if it is possibly related to a drought or flood. Perform a Google search for any recent news items. Call a land manager in the area to inquire of any forest disturbances. Use the “Share this Map” feature of the FCAV and collaborate with other resource professionals, or contact the Eastern Forest or Western Wildland Environmental Threat Assessment Centers to assist with the interpretation.

Why can’t I find my area using the Find Area search feature?

This feature only provides established county names or national forests in the United States.

The Viewer is not responding well, or layers fail to draw.

Performance issues, such as layer drawing, or display refresh issues have been encountered using Microsoft’s Internet Explorer. If you experience negative performance issues with the display of any of the map layers, or masking functions, try using a different browser to run the application, such as Mozilla’s FireFox or Google’s Chrome web browser.

Navigation

Searching

The Map Layers and Map Tool windows

Base Maps