I experimented with applying the LoRa effect to an image of a female high school student caught in a landslide. After acquiring a large number of images of her in the midst of the landslide, I decided to use them to create a scene where a person is almost buried in the debris.
During testing, I noticed that if the person is wearing clothes, the generated image tends to include these clothes prominently. Therefore, if the desired outcome is to have only the face visible amidst the debris, it's better to remove clothing and other accessories from the prompt. A strength of 0.4 is sufficient, but if you want to enhance the effect of being buried in the landslide, increasing the strength to around 0.6 could be beneficial. I also included many images of hands reaching out for help. However, during testing, I found that images often showed distorted fingers, so if you get an image with well-formed fingers, it can be considered a rare find.