Interesting question... a very similar issue was raised by the very recent paper by Efstratiadis et al. (2013, NHESSD):

Efstratiadis, A., Koussis, A. D., Koutsoyiannis, D., and Mamassis, N.: Flood design recipes vs. reality: can predictions for ungauged basins be trusted?, Nat. Hazards Earth Syst. Sci. Discuss., 1, 7387-7416, doi:10.5194/nhessd-1-7387-2013, 2013.

It is freely available here (also comments can be added): http://www.nat-hazards-earth-syst-sci-discuss.net/1/7387/2013/nhessd-1-7387-2013.html

The pessimistic view is no - all we are doing is attempting to tune our models better to the noise. The increase in publications is not a new phenomenon, rather it has been proceding for some time - the problem with things that appear to grow exponentially is that the growth always appears to be a recent occurence. The growth is the logical impact of using publications to measure performance (the publish or perish philosophy).

There have been advances made, not only in our understanding of hydrological processes, but also in our understanding of what we don't know about the processes (e.g. data limitations). However, it is a non-trivial task to dig these out of the literature due to the volume of papers being published.

I agree this is an interesting question. I would suggest you look at this trilogie (a paper / comment and repply to comment)

Wood, E. F., Roundy, J. K., Troy, T. J., Van Beek, L. P. H., Bierkens, M. F., Blyth, E., ... & Whitehead, P. (2011). Hyperresolution global land surface modeling: Meeting a grand challenge for monitoring Earth's terrestrial water. Water Resources Research, 47(5).

Beven, K. J., & Cloke, H. L. (2012). Comment on ‘‘Hyperresolution global land surface modeling: Meeting a grand.

WOOD, E. F., et al. Reply to comment by Keith J. Beven and Hannah L. Cloke on ‘‘Hyperresolution global land, 2012.

The papers and the comment show two approches to modeling. Where Wood et al. aims at increasing the resolution and the complexity of the modeling in priority. Keith and Cloke insiste on the need for new formulations.

Those papers show that still many questions remain unaswered (impact of scale on model formulation for example)

Finaly about the size of publications, lets consider that since we are dealing with stochastic processes a large number of generation (or papers) are needed to converge (to find the Graal).

This appears a basic question that can be posed to many other disciplines, I suppose. So I guess one needs to differentiate between the volume of publications and understanding/improving the science. Although these two are correlated, efficiency and the number of KEY breakthrough articles is another matter. In brief, publications are only one of the indicators for understanding the science; there are many other indicators to be considered as well.

Regarding the rather old knowledge in hydrology, again this could be true for other water related disciplines such as hydraulics, etc., or other geosciences. However, I think first a thorough distinction of what hydrology is all about must be made (differences and overlaps with hydraulics, fluid mechanics, water resources management, etc.). Overall, it is true that papers are rather rare that deal with fundamentals and theoretical aspects of hydrology. Based on my own experience, this is in part attributed to relatively risky and low impact of such efforts/publications. Another major issue is data and cost of data collection systems. Without data, it is hard to make meaningful advances in basic hydrology and one is limited to difficult theoretical advances. There are fortunately new sources of data at basin scale, one of which is remotely sensing. But they still have a long way to go for substituting or effectively complementing the conventional direct data sources; in paricular they still need field data for some sort of training. Thus, you will see many efforts and publications dealing with modeling and uncertainty aspects in hydrology in order to make the best of existing data that are limited in space and time.

The problem is not in the number of publications around already beaten paths. The more are them, the better we will understand the phenomena in hydrology.

The problem is who is able to put them together and to extract the essence from the whole.

This is because nowadays it is hard to have new equations, principles and laws, which have been established well some times ago. Currently, most of the researchers just apply them to explain the phenomena they observed. Not only in hydrology, other subjects also. For example in physics, people still use F = ma, but apply it to some worlds the Newton NEVER knew.

I don't agree with you mr Yang. The low of Newton is demostrated to be very reliable and universally valid in any field in which we have a motion (including the sediment motion, if you want).

The Newton low was developed - see Einstein and further Howking...

Similarly in the hydrology field a simple formula can be complicated as we gain new knowledge about a fenomenon. Having more data and more points of view we can generalize and find the true low of a phenomenon about we only suspect, at this time, to be true.

Starting from Du Boys low, by experiments we have H.A. Einstein new approach and examples can be continued. Because we discuss in a common field what do you think about Parker's lows or C.T. Yang's lows or L. Leopold's lows ? They are contemporary with us...

Applied Hydrology in a catchment or aquifer depends on the interaction of many factors (atmospheric, biotic, lithosphere, energy etc), which are not well understood. While basic hydrological principles like laminar/turbulent flow, condensation, evaporation, groundwater flow have been studied in labs and expressed in analytical or emprirical forms, when these processes interact with soil/geology, climate, vegetation type, etc, to answer qustions for water resource management, uncertainty starts to creep in, and gets huge. besides, most of the world does not have basic hydrological data collected in a standard manner. A simple water balance for a basin is not possible in most of the world, due to lack of longterm data on river discharge, rainfall etc. Evapotranspiration is another huge unknown. To this date there exists no universal way to MEASURE ET for woody vegetation. Models like MIKE and WEAP use formulations with leaf area index and veg height, but a plant ecophysiologist will tell you just how inaccurate these formulations are, if one intends to estimate ET under different plant communities. The only way is to try numerous approaches in different ecosystems, build upon the literature, that can be continually used to improve global datasets like MODIS. Given the crucial importance of water to all life, and the increasing threats upon water resources, a large number of studies, data and pubs are necessary !

Very, very interesting question. In my opinion hydrology is very complex for modelling. There are a lot of models just now, but very few are useful. Thats because the more recient models work with data that are very difficult to obtain. The old models work with more simple data, that are easy to collect. So, in this way, the old models still are very competitive for designing, management and planning. I can use USLE for predicting soil losses in a crop field, just knowing the schedule of field works and vegetal cover along the rotation cycle. Meanwhile, if I want to estimate soil losses with RUSLE, or RUSLE3D, I need to know how organic mater change along the rotation cycle, or how soil rugosity or humidity change along the time.

We need practical models for working every day. Not every worker is a scientist and not have monitored watershed.