Page 189 - Special Topic Session (STS) - Volume 1

STS 423 Avery S. et al.
portions of the JAS questionnaire. Additionally, the statisticians can view other
information layers in CropScape, such as major road networks and the Crop
Frequency Layers for major commodities to guide the manual imputation
process.

Number of Farms
There are regions within the U.S. where pre-screening of small farming
operations is time-consuming, expensive, and subject to misclassification. A
procedure was developed to use the CDL to obtain land cover statistics within
the JAS segments to more accurately identify potential land use, prior to
survey pre-screening activities. The specific goal was to improve the official
estimate for number of farms at the state and national-level for both the JAS
and quinquennial US Census of Agriculture. The most recent five years of the
CDL are used to categorize the land in each JAS segment and estimate
percentages of several predetermined categories, such as percent cultivation,
impervious surface, corn, soybean and pasture calculated at the JAS segment
level. These percentages are used as covariates when modelling the
probability that a record represents a farm and ultimately adjusting the
weights to obtain the estimate of the number of farms for the JAS and Census
of Agriculture.

Natural Disasters
NASS utilizes the CDL and Cultivated Layer to monitor and assess affected
cropland and livestock in the U.S. caused by hurricanes, regional flooding, and
fire events. This capability is now possible due to a refined methodology
utilizing freely available geospatial data products, which include the newly
launched and freely available Synthetic Aperture Radar (SAR) Sentinel-1 data,
optical satellite imagery, and supplemental geospatial hurricane and fire
location data. During disaster response, the confidential in-season CDL data
are used to provide timely crop acreage estimates of impacted land to NASS
stakeholders. The non-confidential (previous year) CDL data are used to
provide crop acreage estimates that can be released within USDA and to the
public on the NASS Disaster Analysis website (USDA/NASS Disaster Analysis
2018).
Specifically for hurricanes and heavy-precipitation events, NASS utilizes
freely available Copernicus Sentinel-1 SAR data to conduct operational flood
mapping of agricultural land in near real-time (Copernicus, 2018). NASS
produces multiple binary inundation raster products that are then overlaid
with agricultural information from the CDL and Cultivated Layer. From this
analysis, NASS is able to estimate the extent of cropland, pasture/hay, and
specific crop types that are inundated. This operational flood monitoring
process provides accurate results based on independent manually-derived

178 | I S I W S C 2 0 1 9