In diffuse noise fields, a superdirective beamformer maximizes the directivity factor at the expense of amplifying spatially white noise in low-frequency bands. It is extremely important to balance directivity factor and white noise gain automatically for practical applications, especially when using large-scale microphone arrays. Considering both environmental noise and spatially white noise, this paper proposes two robust adaptive superdirective beamformers that aim to automatically maximize the amount of noise reduction without introducing a constraint on the white noise gain manually. Specifically, one is an adaptive combination of the delay-and-sum and superdirective beamformers, and the other can be regarded as an adaptively regularized superdirective beamformer. We derive that both the optimal combination parameter and the optimal regularization parameter in each time-frequency bin are related to the power spectral density of the spatially white noise and that of the diffuse noise. Simulations and experimental results show that the proposed two robust beamformers are superior to many state-of-the-art beamformers in different environmental noise and spatially white noise levels.