More direct than this first post that goes more into the process.

Using GLUT data (explanation further down the page):

Forested Acres in 1974 - 174,778 acres or 64% of Hall County

Forested Acres in 2015 (latest available) - 125,343 acres or 46% of Hall County

That's a loss of 49,435 acres of forest between 1974 and 2015. If you simply divide by the 41 years between the two dates you get a loss of 1,206 acres spread across each year.

Using NLCD data (uses different methods than GLUT, limited explanation further down):

Forested Acres in 2001 - 151,986 acres or 55% of Hall County

Forested Acres in 2016 (latest available)  - 142,531 acres or 51% of Hall County

That's a loss of 9,455 acres over 15 years or 630 acres a year.

Tree Canopy in 2011 - 190,080 acres

Tree Canopy in 2016 - 165,568 acres

Pull out the calculator and that is a loss of 24,512 acres worth of tree canopy across Hall County between 2011 and 2016. We could consider tree canopy a measure of green space.

There products from GLUT and NLCD are not directly comparable because they were produced using different methods. The results derived from the methods are clearly different as the values calculated for forested acreage for GLUT and NLCD in 2015/2016 vary by about 17,000 acres. NLCD said that 51% of Hall County was forested in 2016 and GLUT says that 46% was in 2015.

It seems that NLCD data is more than 80% accurate.

Here are the maps again and a little more information:

The oldest available professionally-produced land cover data is from 1974 from the Georgia Land Use Trends (GLUT) which was derived from the LANDSAT satellites that NASA sent up into space that "...provides the longest continuous space-based record of Earth's land in existence." GLUT generated land cover maps interpreting that satellite imagery into distinct classes representing what the land consists of at the time of that photo. Classes include open water, low-intensity urban, high-intensity urban, mixed forest, evergreen forest, etc. The cell size of the imagery was 60 meters in 1974 and 30 meter in 2015. So each pixel in the 2015 photo represents a 30 meter by 30 meter area. See this for more information.

The NLCD data uses 30-meter LANDSAT imagery.

They also released an impervious surface and tree canopy product. You can see the increase in impervious surfaces and decrease in tree canopy.