Temporally variable environmental microseismicity and narrowband signals are both demonstrated to reduce the detectability of small seismic events. We investigate the influence of winter wind events on detection thresholds for a 3-sensor seismic network at the terminus of Taylor Glacier, Antarctica. As wind speeds increase, we observe higher spectral amplitudes across the frequency spectrum; however, some frequency bands are preferentially excited. Surprisingly, these spectral peaks shift frequencies through time. To determine detection thresholds, we implement a waveform injection routine wherein we add scaled waveforms to the datastream, and track changes in the size of the smallest scaled event that we can reliably detect. We thereby demonstrate a capability to quantify the size of the smallest detectable event in temporally variable signal environments. Lastly, we propose a method to forecast our ability to detect sources of a threshold size in measured noise conditions.