Graduate Studies

 

First Advisor

Katherine Frels

Degree Name

Doctor of Philosophy (Ph.D.)

Committee Members

Hugo Gomez, Michael Schlemmer, Peter Baenziger, Stephen Wegulo

Department

Agronomy

Date of this Version

7-2025

Document Type

Dissertation

Citation

A dissertation presented to the Graduate College of the University of Nebraska in partial fulfillment of requirements for the degree of Doctor of Philosophy

Major: Agronomy and Horticulture (Plant Breeding and Genetics)

Under the supervision of Professor Katherine Frels

Lincoln, Nebraska, July 2025

Comments

Copyright 2025, Gerardo Ivan Rivera Collazo. Used by permission

Abstract

Wheat (Triticum aestivum) is an important food staple for many countries around the world and current consumption and production data demonstrate a production deficit. Breeders are challenged to help close this gap in production by selecting better cultivars with improved yields. Several methods aim to accelerate the breeding process to reduce the time required for releasing an elite variety. One of the main breeding bottlenecks is the lack of fast and reliable phenotypic data acquisition that could dissect physiological and morphological plant data. High-throughput remote sensing could have the potential to reduce this bottleneck by streamlining data acquisition, reducing time and effort. To achieve this, a split plot trial was designed with 17 commercially available cultivars as subplots and 10 fungicide treatments applied according to label instructions as main plots. Each pass of plots represented a treatment and the 17 cultivars were randomly distributed in each pass. Correlation analyses by data collection date were performed to evaluate how the sensors data could predict yield variability. Results showed that most of the sensor data was affected by the genetic effect (p < 0.05) with the exception of canopy temperature on 2015 (p = 0.9763). Fungicide did not have a significant effect on most variables, with the exception of reflected photosynthetic active radiation (RPAR p = 0.0074) and fractional photosynthetic active radiation (fPAR; p = 0.005). No cultivar by fungicide interactions were detected for sensor data (p > 0.05). Correlation analysis between sensor data and yield, produced coefficients of determination of up to R² = 0.48 (p < 0.0001) for indices like normalized difference vegetative index (NDVI) during the growth stages between Zadoks 40 and 59. Similar results were obtained for indices such as normalized difference red edge (NDRE), leaf area index (LAI), chlorophyll content (CHL), and adjusted fPAR. Sensor data showed correlations with two hundred kernel weight and plant dry weight. Environmental effects limited the use of sensor based data as predictors of grain protein content. Results suggest that sensor data could become a useful tool to characterize the genetic response of different variables to characterize for yield and yield components that otherwise would require labor intensive or destructive sampling.

Advisor: Katherine Frels

Download

Share

COinS