Improving Crop Estimates by Integrating Multiple Data Sources (2017) / Chapter Skim
Currently Skimming:
3 Multiple Data Sources for Crops: Challenges and Opportunities
3 Multiple Data Sources for Crops: Challenges and Opportunities
Pages 55-78
The Chapter Skim interface presents what we've algorithmically identified as the most significant single chunk of text within every page in the chapter.
Select key terms on the right to highlight them within pages of the chapter.
From page 55... ...
considers multiple data sources when devel oping its county-level crop estimates. The NASS county estimates program produces annual estimates for planted acres, harvested acres, production, and yield (production divided by harvested acres)
|
From page 56... ...
It then turns to the potential use of a new and potentially very important source of recorded data to which NASS currently lacks access -- precision agriculture measurements -- and highlights the benefits and challenges of incorporating this information into NASS's county-level estimation process. This is followed by a discussion of data observed by satellite and other remote sensing tools.
|
From page 57... ...
In 2017, FSA data covered about 95 percent of planted acres in the United States for major commodities. RMA Administrative Data The RMA and the Federal Crop Insurance Corporation (FCIC)
|
From page 58... ...
It also has resulted in the use of a common reporting system for the collection of FSA and RMA data and RMA's adoption of CLU/RLU as its approach for identifying the geospatial location of farm fields. The FSA data account for all farms that participate in FSA programs but not the universe of farms.
|
From page 59... ...
NASS views RMA data on failed acres as more complete than those reported by FSA because farmers must report failed acres to receive insurance payments. While precision agriculture measurements, discussed later in this chapter, are another excellent source of recorded data, they also are unlikely to provide complete coverage of all planted acres.
|
From page 60... ...
Pub. Use Jun JASa, APSb July FSA-578c Aug Acreaged Acreagee Sept APS Acreage Oct Acreage Nov Acreage Dec CAPSf APS Acreage Jan Acreage Feb County estimatesg Mar APS Apr May Jun Yieldh July Aug Sep ARC paymentsi NOTE: APS = Acreage, Production, and Stocks; ARC = Agricultural Risk Coverage; CAPS = County Agricultural Production Survey; FSA = Farm Services Agency; JAS = June Area Survey; NASS = National Agricultural Statistics Service; RMA = Risk Management Agency.
|
From page 61... ...
should collaborate with the Risk Management Agency to obtain relevant individually identifiable acreage and production data and to conduct comparisons with NASS data for the same entity. Future Enhancements to NASS's Use of FSA and RMA Data RECOMMENDATION 3-2: The National Agricultural Statistics Ser vice (NASS)
|
From page 62... ...
This approach of grouping counties with similar type/practice factors is sharply different from the NASS approach of developing estimates based on administrative boundaries, such as county, Agricultural Statistics District (ASD) , and state.
|
From page 63... ...
MULTIPLE DATA SOURCES FOR CROPS 63 FIGURE 3-1 2016 Illinois cash rent paid per-acre nonirrigated cropland. SOURCE: National Agricultural Statistical Service, U.S.
|
From page 64... ...
At a minimum, and relevant to NASS crop estimates, these GPS technologies can provide the GIS boundary of the field planted, area planted, and yield, all by commodity. Precision agriculture measurements are becoming highly valuable.
|
From page 65... ...
farm-level data for a crop. But some of these ground farm-level data now are available through precision agriculture measurements to private organizations that provide decision-making tools and analysis for growers submitting the precision agriculture data.
|
From page 66... ...
RECOMMENDATION 3-5: The National Agricultural Statistics Service should develop a precision agriculture reporting option for the County Agricultural Production Survey/Acreage, Production, and Stocks survey system. Farmers who reported relevant precision agriculture data would either not receive an additional survey form or receive one that was simplified and easy to use.
|
From page 67... ...
These data may be available from other government agencies or for purchase from private companies.7 The current NASS program based on satellite remote sensing and other external sources of information provides indications of planted acres and yield, but only for some crops and some regions. A number of private organizations are now producing satellite-based estimates of yield and publishing them as alternatives to the official NASS monthly estimates of state-level yields during the growing season.
|
From page 68... ...
Climate Divisional Database, which contains a variety of weather and precipitation data;9 Oregon State University's Program in Statistics and Methodology (PRISM) ,10 which provides station-level temperature and precipitation data that may map to CLUs; the Applied Climate Information System, operated and maintained by NOAA Regional Climate Centers;11 and the United States Drought Monitor.12 Multiple Sources of Remote Sensing Data, Sources of Uncertainty, and Models Each of the sources being considered as inputs to modeling has footprints on the earth's surface: • Satellite remote sensing.
|
From page 69... ...
For satellite remote sensing, the natural heterogeneity of pixel contents should be addressed, but likely will be assumed away. NASS Satellite Remote Sensing Indications The methodology used for producing NASS satellite remote sensing indications of planted acres was first described by Graham and Iwig (1996)
|
From page 70... ...
County estimates are constructed by using the coefficients estimated for the state along with the constant term adjusted based on June Area Survey segments in the county. In 2014, satellite remote sensing estimates for planted acres were available for corn in 37 states, for soybeans and wheat in 27 states, for alfalfa in 14 states, and for cotton in 11 states.
|
From page 71... ...
TABLE 3-2 Examples of Classification Accuracy Percentage Percentage Pixels Pixels Pixels Pixels Correctly Actual Incorrectly Incorrectly not Buffered Classified as Classified as Pixels of Classified as Classified as Commodity State Production Pixels Commodity Commodity Commodity Commodity Commodity Corn Iowa 2740.5 million bushels 944,966 397,529 391,624 396,284 1.2 1.5 Corn Ohio 524.7 million bushels 928,450 226,117 220,008 225,195 2.3 2.7 Soybeans Illinois 560.9 million bushels 939,373 280,544 273,966 279,843 2.1 2.3 Soybeans Arkansas 146.6 million bushels 949,805 272,025 264,779 280,643 5.7 2.7 Cotton Texas 7436 thousand bales 1,964,758 222,419 209,102 233,831 10.6 6.0 Cotton Oklahoma 350 thousand bales 943,356 14,729 12,807 15,541 17.6 13.0 Winter Wheat Kansas 4674 thousand bushels 962,256 205,127 196,774 206,806 4.9 4.1 Winter Wheat Minnesota 75 thousand bushels 969,727 276 94 121 22.3 65.9 SOURCE: Based on data from https://www.nass.usda.gov/Research_and_Science/Cropland/sarsfaqs2.php [August 2017]
|
From page 72... ...
and Daytime Land Surface Temperature (DLST) have been demonstrated to be correlated with plant yield during the growing season for a variety of commodities in a number of references, including Johnson (2016)
|
From page 73... ...
. The ICCYF is a crop yield monitoring model that uses remote sensing and agroclimatic data along with survey data to estimate yield during the growing season.
|
From page 74... ...
Hence the model's dependent variable is the crop yield estimate from the November Farm Survey. During development of the model, 80 potential explanatory variables were considered, including same-year yield estimates from the July survey, NDVI (1 km resolution)
|
From page 75... ...
RECOMMENDATION 3-7: The National Agricultural Statistics Ser vice should consider using the 30 percent of Common Land Units withheld to assess the accuracy of classification in the Battese-Fuller regression either as a replacement for or in addition to data from the June Area Survey segments. Estimating Yield RMA data can currently be used to identify insured acres as a proxy for planted acres as well as failed acres by CLU/RLU and production for a farm by crop.
|
From page 76... ...
The increasing number of satellites and the complex types of information they provide make it an ongoing challenge to incorporate the latest satellite data into estimates in a timely fashion. As NASS states in its Estimation Manual,16 "Having the flexibility to use multiple sensors and adapt to new ones is essential in the NASS operational model." Many agencies within USDA are considering how to develop improved production/yield estimates, especially using satellite data and growth models.
|
From page 77... ...
NASS should continue to focus on producing timely and accurate estimates. In light of changing markets and emerging data sources, such as precision agriculture measurements and remote sensing measurements, it should consider whether purchasing private data would support this effort.
|
From page 78... ...
How transparent are the data? RECOMMENDATION 3-9: The National Agricultural Statistics Ser vice (NASS)
|
Key Terms
This material may be derived from roughly machine-read images, and so is provided only to facilitate research.
More
information on Chapter Skim is available.