Not sure how you acquired the data, but unless the machine is also feeding your pace data into training peaks, I don’t believe the pa:hr calculation can be used. It works for outdoor data where you are layering GPS data (pace) and HR, but the recommendation for indoor tests is to set the treadmill at a constant speed and then manually compare your HR drift over the duration of the workout. This is done by comparing the avg HR for the first half (after warm up) to the average HR for the second half and seeing if its in the 3-5% drift. Over 5% drift and your starting HR was too high, under 3% and your starting HR was too low. It appears you are trying to combine both methods into one test.
If the elliptical machine did not track pace and you are relying on your own ability to keep pace at your desired RPM, I wonder if that is accurate enough? I would imagine there are some small fluctuations over the hour that may impact your results. Assuming the pace was 100% stable, the starting target of 153 would be the top end of your Z2 (also tied to your AeT). The 154 and 159 are just to measure the drift, but the HR that matters for the top of the Z2 is the starting HR for the test, so set your top of Z2 at 154 in Training Peaks.