Assuming the waves are shallow-water waves ($d/\lambda < 1/11$) and linear (modeled by a sine or cosine function), the error should be no more than 10% (is that accurate enough? It gets more complicated otherwise).

Since we assumed these are shallow water waves they are nondispersive and the speed is simply $c = \sqrt{gd}$, where $g$ is the gravitational acceleration.

Thus, the wavelength is $$\lambda = c\ T=\sqrt{gd}\ T\,.$$