Yes there is. We can observe things directly in the present. A great deal of natural history is extrapolated. In the here and now, we can repeat an experiment multiple times using the same controls. In the ancient past, you can only find comparable data.
All of science is "extrapolated", and we can dig up bones, and point out telescopes where we haven't before, multiple times. The phrase "comparable data" means nothing at all, as far as I can tell.
You can make educated guesses about the past, but present data is more reliable in general. I think some of your argument confuses between the reliability of data with the reliability of the medium which preserves or disseminates the data.
This is another helping of the same balderdash. Present data is NOT "more reliable". Old data can't be fudged by researchers with an ax to grind, or drunk grad students, or diddled consistently by mis-calibrated instruments, or run over and over until the fingerprints of the results we like causes us to wake up and push the "collect" button. Old data hasn't automatically the taint of pre-conceived notions telling us how to set the instruments next, and what material to feed into the hopper next. New data collection is a very manipulable commodity. You unconscious can sculpt the data as it's laid down. Old data was laid down without conscious, or even unconscious intent. You have the shoe precisely on the wrong foot. If you are a tobacco company, you can decide what data will come into existence. If you are a dinosaur biologist, your data was written down well before anyone had an active iron in the fire about what data would be written down.