Uncertainty is more than a number or colour: Involving experts in uncertainty assessments of yield gaps

Abstract

quality of yield data was ranked as the highest source of uncertainty for actual yields. The justifications provided by experts suggest which uncertainty sources may be reducible with relatively little effort, while other uncertainty sources may be more difficult or impractical to address. SIGNIFICANCE: The decision making on options to improve food production is better informed when uncertainties are accounted for. The proposed uncertainty protocol allows users to distinguish between different sources of uncertainty as well as their level and relative effect on the end result. The ranking of uncertainty sources suggests a prioritization of future effort to reduce the uncertainty around yield gaps. The justifications given by the experts can provide suggestions for options to reduce uncertainty.

The inadequacy of statistical approaches to estimate yield potential and gaps at regional level

“…Second, our analysis focusses only on three crops in the US. We choose the US as a case study because high-quality weather, soil, management, and Ya databases are available and there is large variation in Yw over space for the three crops due to gradients in precipitation and temperature and differences in soil types, providing a suitable testing ground for our comparison 39 . Another source of uncertainty in GYGA can come from the use of single models rather than an ensemble 40,41 .…”

High-Throughput Field Phenotyping to Advance Precision Agriculture and Enhance Genetic Gain, Volume II

“…Parameters are commonly correlated in such a way that their effects on the model output are indistinguishable, leading to what is termed unidentifiability (Cole et al, 2010). Modelling efforts also suffer van Voorn et al 10.3389/fpls.2023.1172359 Frontiers in Plant Science frontiersin.org 02 127 from the existence of multiple candidate model parameterizations and model structures that can describe or explain the data equally well (Beven & Freer, 2001), yet suggest contradictory assessments when focused on practical problems such as yield gap estimates (Schils et al, 2022). High parameter correlations and equifinality are issues that can easily disrupt attempts at an accurate estimation of parameters and thus should be addressed to avoid a reduction in the utility of crop models (Lamsal et al, 2018).…”

A Systematic Review of Geographic Information Systems (GIS) in Agriculture for Evidence-Based Decision Making and Sustainability

“…In order to achieve a precise evaluation of crop production gaps, it is imperative to enhance the quality of input data. This entails obtaining precise meteorological parameters, improving soil characterization, and acquiring geographically dispersed land use data (Schils et al, 2022;Raihan, 2024c). Additionally, the implementation of instrumented geo-referenced validation sites would be necessary in order to provide full survey data that can be used to feed a continuous improvement cycle for assessing yield gaps (Tantalaki et al, 2019;Raihan, 2023j).…”