Skip to main content
U.S. flag

An official website of the United States government

LCMAP - Time of Spectral Change & Spectral Change Magnitude

Video Transcript
Right-click and save to download

Detailed Description

An overview of LCMAP Collection 1 science products: Time of Spectral Change, and Spectral Change Magnitude products.

LCMAP Collection 1 science products include ten science products offering different perspectives and ways to study land change across the country through time. While each product can be used separately, they are complimentary and inform one another.

LCMAP utilizes time series modeling, and an approach known as Continuous Change Detection and Classification, to identify landscape change on a per-pixel basis. All clear Landsat observations are collected over the time period being mapped. CCDC uses a harmonic model fit to the landscape's spectral data. As observations are added to the model through time, the model expects all observations to fit. If the observations diverge from the model, then the model breaks to signify a change. A break requires a new model to be established based on the spectral differences in the new observations. Characteristics of these harmonic models including the timing of breaks and magnitude of breaks, are recorded in the LCMAP products.

While the SCTIME and SCMAG products represent breaks differently, they both describe the same break and therefore are always coincident in time and space. The Time of Spectral Change product is an annual product recording the timing associated with CCDC model breaks. The product provides the day of year that a break was detected in the model. Breaks are recorded as the numerical day of year from 1 to 366, grouped here by month. Pixels with no model breaks are given a value of zero. The LCMAP modeling algorithm requires several observations following a possible break to confirm the break and assign a date. Conditions that reduce observations such as cloud cover may delay the detection of a break, and thus the assignment of a date in the Time of Spectral Change product.

Spectral Magnitude indicates the strength or intensity of the break, identified from differences between observations and model predictions. The SCMAG pixel values are unitless with most having a magnitude of less than 4,000 as indicated by the legend. Again, the zero values are assigned to pixels without breaks.

The Spectral Magnitude and Time of Spectral Change products both illustrate the impact of a forest fire in this region. Time of Spectral Change indicates that the model broke around day of year 235 depending on the individual pixel related to the fire spread. Breaks are detected throughout August, September and October, which align with fire records. The model break in the Spectral Magnitude product was of a low-to-moderate magnitude, which may aid researchers in assessing fire severity.

Model breaks may be associated with changes in land cover, or could result from changes in condition. Some events such as the fire shown here may cause a break in the LCMAP harmonic models, but following the fire, the landscape recovers and retains the same grass/shrub land cover classification.

The Spectral Change products indicate a break in the model but they do not indicate a change mechanism. Users who are interested in the causal factors should investigate additional data sources such as fire records, insect outbreak reports, and urban planning records to further clarify the cause of the break.

The LCMAP Time of Spectral Change and Change Magnitude products allow users to identify the date a change was first detected as well as the relative spectral strength of that change. Used in conjunction with the LCMAP land cover products, these products can provide a wealth of details and information on land surface change, and change in condition. The LCMAP Collection 1 science products can be accessed through Earth Explorer, or the LCMAP Web Viewer. More in-depth information on all LCMAP science products can be found on the LCMAP website.
 

Details

Length:
00:04:12

Sources/Usage

Public Domain.