| dc.description.abstract | The increasing cost of quality cotton seeds, rapidly depleting aquifers used for irrigation, and increasing droughts and extreme weather are making cotton production more expensive and unpredictable in the Texas Rolling Plains and other regions of the world. The research in this thesis explored cotton population density dynamics in irrigated and dryland conditions and tested the utility of a custom remote sensing tool in cotton irrigation management.
The first study was conducted using a replicated split-plot design, with combinations of two irrigation (0 and 90% ET replacement), five population (15,000, 30,000, 60,000, 90,000, and 120,000 plants ha-1), and two plant uniformity (even and uneven plant-to-plant spacing) treatments. The focus was to address gaps in our knowledge related to the impact of plant population and stand uniformity on lint yield and quality in low-yielding dryland conditions. Lint yield was unaffected by plant population and stand uniformity in 2020, but there was an interaction between water inputs and plant population in 2021 in which lint yield increased by 25% from 15,000 to 120,000 plants ha-1 in irrigated conditions only. There was no consistent effect of stand uniformity on lint yield. Fiber quality parameters were minimally and inconsistently affected by population density and stand uniformity.
In the second study, the utility of an all-inclusive sensor system attached unintrusively to a pivot irrigation system, which calculates the Water Deficit Index (WDI) through NDVI, Tc, and in-field weather data, was tested. An experiment was run with four irrigation treatments (0, 30, 60, and 90% ET replacement), and the system was tested in two production fields. The data collected from pivot-mounted NDVI sensors was linearly related to fraction canopy cover (Fc), but the variability and slope of the relationship was altered relative to purely aerial measurements. The WDI values were generally greater than expected, especially before full canopy closure. Integration of a mini weather station into the system was effective in representing on-site weather. The results indicate that the system is quite promising, but more study is needed to better establish the relationship between Fc and NDVI and adjustments to the WDI model are needed to better account for soil heat flux. | |
