Restoration programs operating under high epistemic uncertainty face tradeoffs between spending money directly on restoration and investing in research towards improving management actions. We demonstrate a method to prioritize research needs according to the monetary value of information, based on the potential for additional information to change the optimal amount of habitat restoration estimated by a numerical model. As a case study, we use a population model of the threatened piping plover (Charadrius melodus), which nests on emergent sandbar habitat (ESH) on the Missouri River. Plover reproductive success is dependent upon the availability of ESH, which the US Army Corps of Engineers is managing through habitat restoration and creation due to the rarity of habitat-forming flows in the current hydrograph. The model incorporates uncertainty in the form of estimation error for each parameter. To calculate the value of information, we used the model to determine the level of ESH restoration at which population growth rates will exceed a target with a specified level of certainty. We carried out the analysis using data on ESH availability from before and after the 2011 Missouri River floods, which changed initial habitat conditions and the short-term management actions.
Results/Conclusions
We calculated optimal rates of habitat creation or restoration from the model with all estimation error and with all uncertainty removed from individual parameters. We then determined the reduction in the cost of the optimal management action for each perfectly known parameter value compared to the full-uncertainty scenario. Prior to the flood, we found that information on the average reproductive rate for plovers was roughly twice as valuable as information on first-winter survival, which were the only two parameters for which perfect information led to significant predicted savings in restoration costs. In the post-flood simulations, the large amount of ESH created by the high flows reduced the sensitivity of the plover population growth rate to short-term habitat availability. As a result, the value of information on overwinter survival for both adults and juveniles increased, with potential cost savings becoming greater than those from information on the reproductive rate. While calibrating the method to calculate optimal restoration actions involved some challenges, including limitations on the ability to compare results across different scenarios, we found that the ability to prioritize research needs in terms of potential cost savings increased the utility of the model for resource managers.