Eastern Arc Mountains in Tanzania with cloud forests above 1300 m that constitute the headwaters for many important rivers.

In general...

Native forests in tropical watersheds essentially act as a sponge, by absorbing the force of rainfall and protecting the soil from erosion, by trapping and storing moisture, facilitating percolation of rainwater to groundwater that in turn recharges springs and streams, leading to longer dry season flows ( eg Bruijnzeel 2001, Oyarzun et al 2011 , Krishnaswamy et al 2013 and many others). The precise partitioning of rainfall, percolation, infiltration and evapotranspiration depends upon the forest type, other land cover, climate and soil; however the above processes are generally true.

That native forests promote streamflow in watersheds for the longest duration over the year is seen widely in paired watersheds, with one watershed having intact forests while the adjacent one has significant forest removal for agriculture or seasonal grassland/barren land. The role of forests and water source preservation is an ancient one, with forest protection traditions worldwide (sacred groves) that are reflected today as protected headwater forests, such as those in the Eastern Arc Mountains in Tanzania (pictured above). However, today, unprecedented levels of deforestation has spurred numerous forest hydrology-forest cover studies around the world, with an equally wide variety of results, that have caused confusion.

Confusion on the link between catchment forest cover and water availability

There are many issues in natural resources management where confusion reigns, stemming from the very different perspectives and agendas of various stakeholders. For instance, an often-heard area of confusion is the role forests play in the water cycle. Water managers often believe that cutting forests INCREASES water yield, while reforestation decreases water yields. There is a large body of scientific evidence for this view as well as for the counter-view (eg Hubbart 2009  http://www.eoearth.org/view/article/152819). 

 OR

 "By synthesizing results from studies that used (i) the nested catchment and (ii) the paired catchment approaches, we show that riparian forests decrease water yield on a daily to annual basis."

 - http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3597246/

The main reasoning for increased water yield from deforestation is the reduced transpirative water loss from catchments owing to reduced water uptake by the lesser number of trees present.

However, annual water yield is a VERY different goal from year-round flow in a river. Annual water yield is typically sought after by water resource managers who are concerned with total water availability, for filling large storage reservoirs ( for urban water supply, hydropower generation, irrigation). Whereas local communities that are dependent on streams and rivers for their water require year round flow, and especially flow over the dry months. While removal of forest cover can lead to higher flows in the wet season following rainfall, as well as larger fraction of rainfall as flow because of lowered transpiration losses, the TIMING of flows are altered, to heavier flows in the wet season and very little flow in the dry season. 

Another confusion is the failure to distinguish between fast-growing plantations from native forests. The former usually have much higher water uptake and evapotranspirative losses as compared to native forests, that have evolved with the rainfall and soil moisture regime of the specific area, and in turn, affect the hydrology and streamflow, to which have adapted local aquatic ecosystems.

Studies report on riparian forests depressing the local water table thereby reducing dry season baseflows leading to the drying up of streams, as reviewed by Salemi et al. It is possible that these studies compare streamflow in the presence and absence of riparian tree cover. However, it is to be noted again that greater water availability is not necessarily the best objective when ecosystem conservation ( and preservation of ecosystem services to communities) is concerned. Native riparian forests and adjacent stream ecosystems have evolved over millenia, coupled by water and nutrient cycling. 

References

Bruijnzeel S. 2001. Hydrology of Cloud Forests: A reassessment. Land Use and Water Resources Research 1 (2001) 1.1-1.18

Hubbart, J. (2009). Forest harvest and water yield. Retrieved from http://www.eoearth.org/view/article/152819

David Ellison,*†‡ Martyn N Futter,§ and Kevin Bishop§¶  On the forest cover–water yield debate: from demand- to supply-side thinking.Glob Chang Biol. 2012 Mar; 18(3): 806–820. doi:  10.1111/j.1365-2486.2011.02589.x

Luiz Felippe Salemia, , , Juliano Daniel Groppoa, Rodrigo Trevisana, Jorge Marcos de Moraesa, b,Walter de Paula Limac, Luiz Antonio Martinellia. Riparian vegetation and water yield: A synthesis. Journal of Hydrology doi:10.1016/j.jhydrol.2012.05.061

Jagdish Krishnaswamy a,⇑ , Michael Bonell b , Basappa Venkatesh c , Bekal K. Purandara c , K.N. Rakesh a , Sharachchandra Lele a , M.C. Kiran a , Veerabasawant Reddy d , Shrinivas Badiger a. The groundwater recharge response and hydrologic services of tropical humid forest ecosystems to use and reforestation: Support for the ‘‘infiltration-evapotranspiration trade-off hypothesis’’ Journal of Hydrology 498 (2013) 191–209

Carlos E. Oyarz´un,1* Roberto Godoy,2 Jeroen Staelens,3 Pablo J. Donoso4 and Niko E. C. Verhoest5. Seasonal and annual throughfall and stemflow in Andean temperate rainforests. HYDROLOGICAL PROCESSES 25, 623–633 (2011)