Offshore and onshore wind turbine plant power production, longevity, and operation and maintenance requirements critically depend on site-specific wind speed and turbulence profiles. Thus, understanding of rotor plane conditionsthat deviate from ideal theoretical or modeled values, such as the logarithmic wind speed profile, is important to many aspects of wind turbine performance. The overarching objective of this work is to develop and apply automated quantification methods for non-ideal wind turbine rotor plane conditions. Automated quantification of three major sources of non-ideal wind turbine rotor plane conditions is investigated herein: low-level jets, wind turbine wakes, and wind turbine blade leading edge erosion. The methods presented herein may be utilized within the wind energy field to improve wind resource assessment and dynamic condition monitoring. The occurrence of low-level jets (LLJ) is investigated over the state of Iowa, a state with a high density of wind energy deployment and high wind shear that deviates from classic logarithmic wind speed profile predictions. Low-level jet occurrence and characteristics are also examined over the US East Coast over thirteen of the planned offshore wind energy lease areas that typically experience very low wind shear. Methods to quantify low-level jet wind speed profiles from output from high-resolution numerical weather prediction models are evaluated. State-of-the-art low-level jet quantification algorithms are shown to exhibit sensitivities to wind speed profile vertical resolution, and methods that reduce these sensitivities are presented. Offshore low-level jets are associated with particular meteorological conditions such as highly stable atmospheres and low boundary-layer heights, and present with distinct rotor plane characteristics when compared to ideal wind speed profiles (i.e., the logarithmic wind speed profile). Further research on using state-of-the-art image processing machine learning models is applied to the problem of quantifying wind turbine wakes measured at large wind farms. Scanning lidar measurements are now able to accurately determine wind speeds over more than 10 km distance, covering dozens of wind turbines. To automate the processing of high-volumes of scans, a convolutional neural network (CNN) image processing model is developed and trained. The developed models produce high accuracy wake quantification over various lidar scan configurations and atmospheric regimes, and they generalize well to lidar scans of wind turbine wakes collected in complex terrain, offshore, and near a turbulence-enhancing, near-shore escarpment. Finally, the methods developed are further matured and expanded for quantification of leading edge erosion (LEE) from field images. Both unsupervised and supervised (CNN) methods are utilized and both are able to detect over 60% of pixels in an image that were classified as damage. Accuracy in distinguishing between shallow and deep leading edge erosion damage is exhibited, and the extent of the blade covered in shallow or deep damage is automatically quantified. Overall, the models developed have been applied to a number of wind energy cases and are able to detect and quantify non-ideal rotor plane conditions attributed to LLJ, wakes, and blade LEE. These types of models are critical to moving forward the automated detection and quantification of specific issues relating to wind energy in large measured and modeled data volumes. Additionally, for automated LEE and wake quantification, the models could also be applied for real time condition monitoring (i.e., image processing of lidar wake scans for wake steering purposes or image processing of wind turbine survey images collected by drones for blade maintenance planning)