The Swerling models quantify the signal loss that a radar experiences because of fluctuations in the power reflected by a target as its orientation relative to the radar changes with time. See the Wikipedia article "Fluctuation Loss" for a discussion.
It is not my field directly, but maybe this helps:
Abstract:
In this work, we study the fundamental tradeoff between integration time and scan rate in coherent radars. The contrasting needs for a large probability of detection and a short scan time are carefully balanced by optimizing the detection rate, defined as the average number of detections from a target per unit of time. A solution for the optimum pulse train length is provided under Swerling 0 (non-fluctuating), Swerling I (χ 2 with two degrees of freedom), and Swerling III (χ 2 with four degrees of freedom) target fluctuation. Some examples are given to show the possible tradeoffs among the principal system parameters.
Published in: 2017 IEEE Radar Conference (RadarConf)
Swerling target models are a set of statistical models used in radar signal processing to describe the statistical behavior of radar echoes from targets.
The Swerling models are named after Peter Swerling, who developed them in the 1950s. They are based on the assumption that a target's radar cross-section (RCS) fluctuates randomly as it moves relative to the radar, due to changes in its orientation, shape, and size.
There are five Swerling target models:
Swerling 0: A non-fluctuating target model where the RCS of the target remains constant over time.
Swerling 1: A fluctuating target model where the RCS of the target varies randomly over time according to a Rayleigh distribution. This model is often used to represent a target with multiple scattering centers, such as an aircraft.
Swerling 2: A fluctuating target model where the RCS of the target varies randomly over time according to a Chi-squared distribution. This model is often used to represent a target with a single dominant scattering center, such as a missile.
Swerling 3: A fluctuating target model where the RCS of the target varies randomly over time according to a mixture of Rayleigh and Chi-squared distributions. This model is often used to represent a target with a combination of scattering centers, such as a ground vehicle.
Swerling 4: A fluctuating target model where the RCS of the target varies randomly over time according to a Gaussian distribution. This model is often used to represent a target with a highly variable RCS, such as a small boat on the water.
Each Swerling model represents a different type of target and has a different statistical distribution for the target's RCS. Knowing the Swerling model for a given target can help radar designers optimize their systems for detecting and tracking that target