First off, here are some basics about light measurement. The PAR sensor measures light in micromoles per square meter per second. What does it mean? Simply put, it counts the number of photons – particles of light – that fall onto the detector surface each second. The detector surface is flat, so the incident light angle is crucial as well. When the sun is at its zenith (directly overhead), the sensor receives its light entirely. If radiation comes in at an angle, the sensor receives less light.

My PAR sensor curve is jagged. Shouldn’t it be smooth? What is happening?

There are two factors that can make PAR curve jagged.
The first one could be caused by incorrect placement. If the sensor is placed in a greenhouse structure, this casts a shadow on the PAR sensor and the curve will drop.

The second one is more subtle. At certain times in the morning and evening, there are jumps in the measurement. It seems that the detector suddenly receives more light. Why does it happen?
This is due to the clouds, which might initially seem counterintuitive. When the morning sunlight hits the detector at an angle on a cloudless day, the radiation received follows the cosine law.
If there are clouds right above the detector, solar radiation can be reflected onto the sensor by the clouds, even if they do not block the sun. In this case, the detector receives the same amount of sunlight as in the first case, plus the light the cloud reflects. As a result, the PAR reading is higher.