If that's your typical exponent, you can use some trivial math to get around this.
Rather than feeding in the negative value, send in the absolute value. This will give you 2009.05 in this case. Once you perform the calculation, you can use negate to give it back the correct sign. Depending on your input, you may need to add a little logic in the front to send positive values through cleanly and negative as their postive counterparts. But, it'll let you perform the math you're looking for.
I've looked for other power options using a negative x and they all result with the same NaN. It's possible someone else will be able to find a more elegant solution. But, I did test this one out to see that it works. As you were hardcoding the exponent, I'd expect this to work. Obviously, if you were going to get into something like a square you'd have to do something a bit more tricky to determine which sign should apply.
