ROS Resources: Documentation | Support | Discussion Forum | Index | Service Status | Q&A answers.ros.org
The results of the inverse transform after rosrun tf tf_echo base_footprint utm are actually correct. When you invert a transform, the translation component isn't simply negated. Rather, the translation component of an inverse transform is:
Also, the reason you are getting different results for the Euler angle test is because you are inputting a roll angle instead of a yaw angle. The sequence of the Euler angle arguments to the static_transform_publisher is yaw, pitch, roll.