India's plans for the Inter Basin Water Transfer Project represent one of the most ambitious efforts yet proposed to engineer a country's hydrological system. The ecological impact of the proposal on fish communities has not been previously studied despite its potential for serious and irreversible changes to the biodiversity and biogeography of India's rich native freshwater fish community. Such a large scale manipulation of river network geometry represents an unusual opportunity to apply the theory of dendritic ecological networks to an empirical system which is of both academic and conservation/management interest. We apply a neutral metacommunity model to scenarios representing both simulated and empirical data to show how freshwater fish communities are likely to be impacted by this project. We also consider what changes to the current plan would mitigate potential disturbance to the fish fauna in order to create general principles of design that satisfy plan goals (redirecting water from water-rich to water-scarce areas) while minimizing impact on fish communities. 
Results/Conclusions

We find that, when using simulated datasets for the spatial distribution of freshwater fish, the addition of the manmade canals facilitates the spread of common species at the expense of rare species, allowing them to become even more widespread. In many scenarios, this lowers the total number of species in the system. This is of particular concern when considering those Peninsular links which connect the smaller, endemic-rich river basins west of the Western Ghats with the much larger river basins to the east. We also find that the river linking project causes local fish communities to be, on average, less similar because of the canal-facilitated importation of new species from distant river basins on the Peninsula. The extent to which fish communities are affected is strongly influenced by the dispersal mechanisms of the fish, a factor which must be fitted from the site-specific empirical data available.