Firstly, I think it would be a good idea to put on a thinking cap and formulate a research question and a testable hypothesis, then let the question/hypothesis guide your choice of statistical method. Using the data to formulate a hypothesis and then testing the hypothesis on the same data is usually a bad idea -- it might appear to give good results, but you would be in danger of attributing a lot of meaning to statistical noise.

Anyway, if you are just exploring the data, I think a good place to start would be to perform a two-sided t-test on each of the parameters individually -- that should give you a rough initial idea of which variables differ significantly between protected and unprotected areas.

Since you have many variables, I think a dimensionality reduction technique such as Principal Coordinate Analysis (PCA) might be a good starting point too. It is a very useful way to visualise your dataset and pick out the most significant characteristics.

Most statistical textbooks will give you the low-down on both the t-test and PCA.

But seriously, if you don't formulate a hypothesis first, then at least see if you can divide your dataset in two samples: one sample for exploration/estimation, and another sample for out-of-sample validation.

Firstly, I think it would be a good idea to put on a thinking cap and formulate a research question and a testable hypothesis, then let the question/hypothesis guide your choice of statistical method. Using the data to formulate a hypothesis and then testing the hypothesis on the same data is usually a bad idea -- it might appear to give good results, but you would be in danger of attributing a lot of meaning to statistical noise.

Anyway, if you are just exploring the data, I think a good place to start would be to perform a two-sided t-test on each of the parameters individually -- that should give you a rough initial idea of which variables differ significantly between protected and unprotected areas.

Since you have many variables, I think a dimensionality reduction technique such as Principal Coordinate Analysis (PCA) might be a good starting point too. It is a very useful way to visualise your dataset and pick out the most significant characteristics.

Most statistical textbooks will give you the low-down on both the t-test and PCA.

But seriously, if you don't formulate a hypothesis first, then at least see if you can divide your dataset in two samples: one sample for exploration/estimation, and another sample for out-of-sample validation.

In addition to the suggestions of Ian and Alessandro, you may also consider time-series analysis of the data as this statistical method has been used extensively in studies pertaining to vegetaion phenology. Below you find some suggested literature based on the time series analysis for vegetation phenology:

http://www.rali.boku.ac.at/fileadmin/data/H03000/H85000/H85700/publications/21_AtzEilers_2011_IJDE.pdf

http://www.sciencedirect.com/science/article/pii/S0034425705001057

http://link.springer.com/chapter/10.1007%2F978-90-481-3335-2_9#page-1

If you run into article accessibility issues, please let me know. Moreover, some studies also suggest the usage of geo-statistical interpolation technique. Nevertheless, as Ian suggested - the first thing is to refine the scenario for your research and its objectives and then different tools can be assessed and compared based on their suitability.

hope this helps!

Thank you very much for the suggestions, much appreciated.