Activity Number:
|
615
- Bayesian Methods for Complex Survey Designs and Data
|
Type:
|
Topic Contributed
|
Date/Time:
|
Thursday, August 3, 2017 : 8:30 AM to 10:20 AM
|
Sponsor:
|
Section on Bayesian Statistical Science
|
Abstract #324395
|
|
Title:
|
A Model-Based Approach to Crop Yield Forecasting
|
Author(s):
|
Nathan Cruze* and Habtamu Benecha
|
Companies:
|
USDA-NASS and USDA National Agricultural Statistics Service
|
Keywords:
|
Bayesian model ;
model-based estimation ;
agricultural statistics
|
Abstract:
|
The USDA's National Agricultural Statistics Service conducts multiple surveys for major crops during the growing season. These surveys are designed to capture the current status of crops at state, regional, and national levels with a first-of-the-month reference date. Each of the surveys also provides an estimate of potential end-of-season crop yield. We extend a Bayesian hierarchical model to produce improved yield forecasts for upland cotton. The model combines these possibly disparate survey estimates together with auxiliary data to produce one-number forecasts for a region and its member states. The resulting state forecasts are benchmarked against the regional forecast. The model gives rise to easily reproducible estimates with rigorous measures of uncertainty. The proposed candidate model for upland cotton is shown to perform well over a wide variety of growing conditions. Some particular challenges of modeling upland cotton are noted.
|
Authors who are presenting talks have a * after their name.