This work is a U.S. Government work and is not protected by U.S. copyright law. It was created as part of the author's official duties. The dataset presents 2100 Landsat Thematic Mapper (TM) and Enhanced Thematic Mapper Plus (ETM+) surface reflectance scenes, providing 30-m resolution coverage for North America for epochs centered on 1990 and 2000. The data were processed using the Landsat Ecosystem Disturbance Adaptive Processing System (LEDAPS) at NASA's Goddard Space Flight Center. The raw data were obtained from the orthorectified Landsat GeoCover dataset, purchased by NASA from Earth Satellite Corporation. The LEDAPS project calibrated and atmospherically corrected the data using the MODIS/6S methodology. Initial comparisons with ground-based optical thickness measurements and MODIS imagery indicate comparable uncertainty in Landsat surface reflectance to the standard MODIS product. The dataset supports decadal assessments of environmental and land-cover change, production of reflectance-based biophysical products, and applications that merge reflectance data from multiple sensors. The LEDAPS project reused the MODAPS software architecture for producing higher-level products from MODIS Level 1B data. The processing involved ingesting and calibrating Landsat GeoCover data, correcting to top-of-atmosphere (TOA) reflectance, and applying atmospheric correction using the 6S radiative transfer code. The dataset includes aerosol optical thickness (AOT) estimates derived from the dark, dense vegetation (DDV) method. The TOA and surface reflectance maps preserve the 30-m resolution of the original GeoCover data. The dataset also includes a 500-m resolution product derived by aggregating the 30-m reflectance data. The LEDAPS surface reflectance dataset supports multiple applications, including land-cover mapping, decadal land-cover change, surface water resources, and vegetation biophysics. The conversion to reflectance allows users to cross-compare Landsat observations to laboratory or ground-measured spectral curves, reflectance data from other instruments, or the output from canopy reflectance models. The dataset has been validated against MODIS and AERONET data, showing reasonable agreement. Future work includes extending the dataset to the 1975-era MSS GeoCover dataset and correcting for adjacency effects in AOT maps. The LEDAPS project demonstrates the potential for fully operational atmospheric correction of Landsat-type imagery, supporting routine land-cover assessments every five years.This work is a U.S. Government work and is not protected by U.S. copyright law. It was created as part of the author's official duties. The dataset presents 2100 Landsat Thematic Mapper (TM) and Enhanced Thematic Mapper Plus (ETM+) surface reflectance scenes, providing 30-m resolution coverage for North America for epochs centered on 1990 and 2000. The data were processed using the Landsat Ecosystem Disturbance Adaptive Processing System (LEDAPS) at NASA's Goddard Space Flight Center. The raw data were obtained from the orthorectified Landsat GeoCover dataset, purchased by NASA from Earth Satellite Corporation. The LEDAPS project calibrated and atmospherically corrected the data using the MODIS/6S methodology. Initial comparisons with ground-based optical thickness measurements and MODIS imagery indicate comparable uncertainty in Landsat surface reflectance to the standard MODIS product. The dataset supports decadal assessments of environmental and land-cover change, production of reflectance-based biophysical products, and applications that merge reflectance data from multiple sensors. The LEDAPS project reused the MODAPS software architecture for producing higher-level products from MODIS Level 1B data. The processing involved ingesting and calibrating Landsat GeoCover data, correcting to top-of-atmosphere (TOA) reflectance, and applying atmospheric correction using the 6S radiative transfer code. The dataset includes aerosol optical thickness (AOT) estimates derived from the dark, dense vegetation (DDV) method. The TOA and surface reflectance maps preserve the 30-m resolution of the original GeoCover data. The dataset also includes a 500-m resolution product derived by aggregating the 30-m reflectance data. The LEDAPS surface reflectance dataset supports multiple applications, including land-cover mapping, decadal land-cover change, surface water resources, and vegetation biophysics. The conversion to reflectance allows users to cross-compare Landsat observations to laboratory or ground-measured spectral curves, reflectance data from other instruments, or the output from canopy reflectance models. The dataset has been validated against MODIS and AERONET data, showing reasonable agreement. Future work includes extending the dataset to the 1975-era MSS GeoCover dataset and correcting for adjacency effects in AOT maps. The LEDAPS project demonstrates the potential for fully operational atmospheric correction of Landsat-type imagery, supporting routine land-cover assessments every five years.