Given arguments upon criticism of an anonymous peer-review of the paper submitted to the Rus-sian "Water Resources" seem to clarify our brief report on radar. Noticeably is that review has been done by a meteorologist which is not versed in engineering hydrology, therefore these argu-ments are to address its special issues once again. Let us name our arguers as Atmos and Hydros (you should guess who are they) and imagine the discussion between them.

Atmos: How accurate are the radar data obtained at Valday? How much may become the er-rors for remote areas off the radar site?

Hydros: The accuracy problem was beyond the scope, but I could refer to a lot of works dedi-cated to the problem. Raw radar precipitation data were previously calibrated by rain gauge network. The calibration consisted of comparison of the radar and ground-based precipitation amounts. Evaluated are either the biases gauge-radar (G - R) or the ratios G/R. The latter characteristic is highly variable: from 0.7 to 2.2 for daily rainfall amounts. It generally corre-lates with the distance off the radar and, slightly, with the land surface relief.

Atmos: You wrote that radar gives a possibility to specify the extreme rainfall rates. In com-parison to what it does? To the rainfall gauges or rain recorders (pluviographs)?

Hydros: It specify considerably the extreme rates for a river drainage basin. It is quite obvious any gauge network is discrete, often sparse on given watershed or even absent at all, so the areal extremes are determined arbitrary. Also, the river watershed responds to a rainfall differ-ently when the rain covers either its headwaters or the outlet. Hence, not only extreme rainfall rates should be taken account, but where they are concentrated.

Atmos: About the areal reduction (AR) of rainfall rate. Does it define as in your paper, and would you mind to improve it later on. I think their values still are erroneous as normalized by a local extreme which may be measured incorrectly.

Hydros: Since the AR value is determined as the ratio of averages to a local (or "focal" if ex-actly) extreme, it can disproof any climatic issues, but only that value, its dynamics may pro-vide an input to a flood model. Given the extreme as a mean, climatically approved value, we can get a paradox: AF > 1, when unexpectedly heavy rain would occur. And, those dynamic extremes and the respective ARs would attribute each of the storms to be classified by them-selves. Earlier, our prominent specialist Dr. Livanova have obtained (and I proved them again by using more fresh data) two different empirical equations for AF, one for the frontal rains, another for the air mass (convective) storms.

Atmos: What about the "pulses" or "flux of events" you mentioned there? I do not know what they are: one can hesitate to sense what is the "flux". Maybe, it is the "flux of impulse", that often termed abroad as "momentum transfer" in atmospheric surface layer?

Hydros: It is a term from the so-called stochastic hydrology, that branch of science which ex-amines the river runoff time series as the random process. Under "flux of events" we assume here the series of those rainfall events, when rainfall rates exceed some threshold. The latter we suggest to be defined by antecedent moisture (infiltration capacity) of the river drainage basin. The "flux" of such "pulses" comprises both determined component and random noise. The magnitudes can be or independent of each other, or be in some "tele-connection" that still remains an arguable issue. To the simplest, the first case we may (and did so) fit the Poisson distribution for frequency of the events and spaces among each other.

Atmos: You wrote "the temporal reduction (TR) of precipitation rates depends on the run time". But, it is the ratio of extreme and mean (timely averaged) rainfall rates. Is not it? And, neither the extremes nor the means cannot depend on any runoff times, as are both only at-mospheric dependent phenomena.

Hydros: There is again a misunderstand between the atmospheric and water scientists. As proffered by engineering hydrology manuals, the TR-factor is determined as the ratio of ex-treme rainfall observed for the run time (individual to a river basin) to daily averaged rate. The long is the run time (e.g. for a large basin), the higher rates are likely to be observed (in prob-abilistic sense). The manuals contain approved relationships "TR - run time". So, for our con-ditions of the hilly uplands of north-western Russia, we may adopt the following:

These values are to be (and being now) specified for several basins within a radar reach.

Atmos: As it seems, you should (but did not, why?) more explicitly define the tasks. The chapters 4.4 and 4.5 of your work, dedicated to the "conventional" probability and classifica-tion of rainfall fields, respectively, do not comprise any information special of the examined problem. Both they are irrelevant to runoff computation with the use of radar data.

Hydros: I cannot agree. Engineering hydrology is not to only calculate the flood runoff as probabilistic, but allows to support the water management, water conservation and some envi-ronmental issues. One of them is how to evaluate the areally distributed (potential) runoff, it becomes challenge when we are going to design flood for an extensive area, e.g. along a main pipeline route. The newly obtained conventional probability function is to detect how much is the rainfed area during a steady rain of given rate and frequency. Or, conversely, how heavy may be that rainstorm covered given area with a runoff depth of certain magnitude. Remem-bering about the time series, we should developed a model for "time series of rainfall fields". We have to (and made an attempt to) classify the rainfall fields (spatial distributions) obtained from radar observation for a long time. Having the time series of the types, one can explore their frequencies, conversions to one another and so on, that can promote for climate-related investigations.

Atmos: You wrote "a basis has been proposed for rainfall classification as applied to climatic research". I think it is claimed too strongly.

Hydros: I think just so. Therefore, let's go to make such a basis to be strong.

Hydrosphere © 1990 — 2005 Vladimir A. Shutov
Design by Timothy (Azs) Shutov © 2001