This is an interesting question for me because my first research internship centered pretty much entirely around this. It got pretty deep into the weeds as this was a CS, rather than biology, internship and we were mostly analyzing the math behind the allometry (e.g. Log-transformed linear vs. nonlinear regression for fitting the curve, OLS vs. RMA regression, traditional vs. monte carlo methods for comparing curves with different numbers of parameters, etc.) but much of the work we did is applicable here.
So, what you have here is DBH, or Diameter at Breast Height, which, lucky for you, happens to be the most common method for estimating the biomasses of trees. Here is the way that it works. Basically, a research team will go out and measure the diameter of a bunch of trees. Then they will cut down those trees and weigh them. What they will end up with is a graph with a bunch of points that give you the weight of trees of particular diameters. Then you fit a curve to this graph and, next time you want to estimate the biomass of a tree, you plug that tree's diameter into the equation for your curve. There are a bunch of databases containing allometric equations for different tree species. You just need to find one that contains the trees that you are interested in. Given that you are working in New York, this might have what you need:
Though you are better off figuring out the species of the trees you are looking at and seeking out allometric equations for those specifically. There might even be allometric information included with the data you are being provided. One thing that may be an issue is that correlations between DBH and biomass are much more common than correlations between DBH and height. Though both exist, you may have more trouble tracking down the latter for specific cases. Additionally, these numbers are very rough estimates that vary a great deal depending on the techniques used to collect the data, fit the curve, and extract the equation. In my own research, I found that log-transformed linear regression tended to be more successful over all (oddly enough) but that nonlinear regression was more effective on young forests dominated by stands of small trees. We also had a rather odd finding that log-transformed linear regression of two lines with a breakpoint tended to be less accurate in both situations but more accurate overall (i.e. it did better on big trees than nonlinear regression and better on small trees than linear regression) but we weren't able to come up with a mechanism to explain why this would be the case so we scrapped the findings. At any rate, all of this should illustrate that these numbers are very rough estimates that can be heavily influenced by environmental biases and that different research teams may even disagree on the proper allometric equations to use on a given species. This is why databases are helpful, as they tend to average out a large number of allometric studies rather than relying on a single one, but even these should be taken with something of a grain of salt.
If you are expected to actually come up with these equations yourself, then you need to actually measure the diameter and heights of a bunch of trees, plot them, and fit a curve to them. But from the sounds of it, you are only expected to estimate tree height based on diameter. If that is the problem you are being asked to solve, chances are that there are allometric equations available to help solve it, perhaps even ones that were provided to you.