A key task in Wireless Sensor Networks (WSNs) is to deliver specific information about a spatial phenomenon of interest. To this end, a few Sensor Nodes (SNs) sample the phenomenon and transmit the acquired samples, typically multihop, to the application through a gateway called sink. Many applications require the spatial sampling to be accurate and the delivery to be reliable. However, providing a higher accuracy/reliability comes at the cost of higher energy overhead as additional messages are required: increasing the number of samples to increase the accuracy of sampling and increasing the number of retransmissions to increase the transport reliability. Existing design approaches overlook optimized spatial sampling accuracy and transport reliability in combination for minimizing energy consumption. This work aims in providing the optimized solution for sampling accuracy and transport reliability in composition for a maximized efficiency. Our approach features a message efficiency that optimally meets application requirements with the online adaptation and appropriate tradeoff between accuracy and reliability. The sampling and transport co-design proceeds by finding optimal number of SNs for the accuracy of the spatial sampling with the effect of reducing the number of retransmissions and still satisfying the application requirements. We validate the approach viability through analytical modeling and extensive simulations for a wide range of requirements. © 2013 IEEE.