Random access schemes are primarily used for data transmission in the uplink of cellular networks. Every user in a random access network is programmed to follow a predetermined transmit control policy that is designed to achieve optimal network performance. This approach, however, is not very efficient in cellular networks where channel conditions vary from time to time. Employing a fixed transmission policy may not guarantee optimal performance. To alleviate this problem, recently, channel aware random access schemes have been proposed wherein information available at the physical (PHY) layer is utilized at the higher layers to maximize network throughput. Such a cross-layer approach naturally has its share of challenges and problems. The objective of the proposed research is to study the effect of multicarrier diversity on channel aware random access schemes. First, we describe two generic random access schemes - channel aware multicarrier random access (CAMCRA) and no selection random access (NS-RA) for multicarrier networks. The former is based on judicious carrier selection and exploits multicarrier diversity while the latter does not perform carrier selection. For illustration purposes, we consider the well-known Aloha protocol and study the effect of channel state imperfection on the overall network throughput. We show that networks employing the NS-RA scheme are extremely sensitive to channel measurement errors. More precisely, the asymptotic average throughput of the NS-RA scheme under uncertain channel conditions is zero. The CAMCRA scheme, however, is very robust to channel estimation errors and maintains the same order of throughput.