Printer-friendly Version
Contents
1. conclusions based on the simulations are not given in much detail Even with one year of data some interpretation can be made regarding further modification or addition in the system A We have added the appropriate parts in the revised text better explaining the meaning of the simulation and enhancing the conclusions Q 6 Model details especially the ADIGE components are given in detail but they do not highlight what is new different from those used before So in my opinion details of the model should come first then a comparison can be made with others and not vice versa A What is different from the previous model is the assembly of the various model C654 components The single parts used to describe sub processes were in fact derived from literature and not invented here However the main novelty is that these components can be more easily interchanged than in the past by just introducing the appropriate component and linking the new one to the others at run time as opposed to compiling time This capability endows the JGrass NewAGE system of great flexibility and allows it to grow with knowledge of processes and basins without the need to rebuild the whole code We thought the GMD the appropriate journal for the dissemination of this achievement since it is said to be an international scientific journal dedicated to the publication and public discussion of the description develop ment and evaluation of nume2. C73 w3 1 4 3 fas The left hand side of the equation 1 is expressed by the derivative of the equation 5 After some mathematical steps the equation 6 gives the non linear ordinary differential equation in the unknown Q dQi t dt K Qilt Qgen t E Qro Qi t i 1 2 H 6 trib The coefficient K i t is equal to ire tied PE RE g C668 Ko NI w where C L 3 T is the Chezy coefficient b L and L represent the width and average length of the link respectively i is the average slope of the link and Q L T t is the channel discharge For a more detailed discussion of the terms in Eq 7 see Menabde and Sivapalan 2001 and Mantilla et al 2005 We hope that the concepts results more clear now Q 7 p 952 eq 14 The definition of Chezy coefficient seems to differ between eq 13 and 14 Qsp is not defined A In the revised text we have better specified the meaning of the variable Q 8 p 952 I 16 How the hypsographic curve is used in the model is not clearly explained A The hypsographic curve is not used in the model We have removed this part because it is not relevant in this application Q 9 p 952 Il 19 20 Please provide the details about the Passer river basin that is used in this analysis A We accept the suggestion in fact we have provided a short description of the main hydrological feature of the new basin Little Washita in the rev
3. and Rodriguez lturbe l Geomorphological theory of the hydrological response Hydrological processes 10 803 829 1996 Rinaldo A Marani A Rigon R et al Geomorphological dispersion Water Resour Res 27 513 525 1991 Rinaldo A Vogel G Rigon R and Rodriguez lturbe l Can one gauge the shape of a basin Water Resources Research 31 1119 1127 1995 Rodriguez lturbe and Rinaldo A Fractal river basins chance and self organization Cam bridge Univ Pr 2001 C674 uez1992energy Rodriguez lturbe l Rinaldo A Rigon R Bras R Marani A and Ijjasz Vasquez E En ergy dissipation runoff production and the three dimensional structure of river basins Water Resources Research 28 1095 1103 1992 Saco P and Kumar P Kinematic dispersion effects of hillslope velocities Water resources research 40 W01 301 2004 Slatyer R and Mcllroy l Practical Microclimatology Practical Microclimatology 1961 Snell J and Sivapalan M On geomorphological dispersion in natural catchments and the geomorphological unit hydrograph Water Resources Research 30 2311 2323 1994 C675
4. for similar trials which would need research of its own Q Page 951 Line 11 How do you determine the residence times for each hillslope What are the controlling factors Heading section 3 use a large P for river Passer A We have changed model and we no longer use this scheme However residence time for each hillslope was determined using the geomorphic path ways determined as in D Odorico and Rigon 2003 and jgrasstools modules http code google com p jgrasstools Q Page 952 line 14 north east instead of nord est A Corrected C647 Q Line 15 25 It is not explained why you derive and present all this in formation describing topographical features Where is such information used in the model Please explain this or consider to omit it as irrelevant information A Because we decided to change the river basin on which the model the question was applied this is no longer relevant By all means we had added that information solely to describe the topographic features of the basin and they were not used in the model In the new version this topographic information is not reported Q Page 953 line 1 3 Why do use such short periods of only a couple of months Is that a data problem A The main problem was to avoid snow generated discharges which would have involved the need to explain in detail how the snow model works In the new draft the model has been applied to the Little Washita river basin
5. manuscript so reducing the number of parameters Q Page 950 line 16 Ks in Darcy s law is the saturated hydraulic conduc tivity do you adapt it due to different saturations in your hillslope Or do C646 you always assume saturation prior to runoff generation What about e g preferential flow in macro pores A No we did not But we could We used a single calibrated mean value for the whole catchment If flow in macro pores is Darcian then their contribution would raise the calibrated value However the Hymod runoff model uses different concepts Q Line 18 As Ks is calibrated wonder whether you use distributed pa rameters i e different values on different hillslopes or do you have only one parameter for the entire model catchment If yes how do you deal with different physical soil properties Some more explanation on this would be interesting A No we used one Ks value which is constant throughout the Passer basin and equal to the value obtained from the manual calibration Having more gauges we could calibrate the different parts of the basin differently It is true that in principle we could have used pedotransfer functions or similar to associate an estimation of the hydraulic conductivity at saturation to any soil type and or land use However soil maps are not very reliable for Passer at least for hydrological use and the topics raised by their use require further care see for instance Terribile et al 2011
6. quantities is not very in formative please be more precise what you mean A The sentence under review is Models developed to reproduce a whole set of hy drological quantities for operational purposes came from water resource management and agriculture needs The sentence has been changed to The water resource and river management required however the need to estimate a whole set of hydrological quantities such as discharge evapotraspiration soil moisture bringing very soon to the implementation of more comprehensive modeling systems as the pioneering Stanford watershed model Crawford and Linsley 1966 or like the Sacramento model e g Burnash et al 1973 the PRMS model Leavesley et al 1983 Q Line 29 Please indicate how you define an acceptable degree of con fidence C641 A The sentence under review is A solid paradigm of simpliifAAcation is offered by the theory of the geomorphological unit hydrograph which provides ow values at a single point of the river network i e at the outlet of the basin In this case many models with few parameters are able to reproduce the expected result with an acceptable degree of coninAdence The sentence was removed in the last version of the paper Q Page 946 line 2 3 I do not agree with the statement the topology and the geometry of the river network is more important A The sentence is e g Rinaldo and Rodriguez ltur
7. the revised text about what happens if some of these tools are excluded and we think this can be considered besides the whole thing itself a further original contribution We believe that in doing this we partially fill the gap between writing of scientific soft ware and commercial practice which profitably uses the concept of object oriented programming Again object oriented programming is nothing new but it has proven to be a successful tool in solving many issues connected with cooperative work of many As objects models algorithms and data can be packaged in components they can expose for reuse only their most important functions Libraries of components can then be re used and efficiently integrated across modelling efforts Therefore the modeling experience can exploit properties such as encapsulation data abstraction and inheri tance which can greatly help towards better science at least when code writing is the method We think that the components presented in this paper can be improved with method being able to single out the part of code responsible when simulations do not agree with computation as we have done in substituting Duffy s submodel Besides we C665 think we have given hydrologists a flexible tool which allows them to spend more time understanding the physics of the processes rather than in other tasks Q 2 p 949 eq 3 6 Explaining the underlying physical concepts in more detail here in
8. 760 millimeters and temperature of 16 degrees Celsius winters are typically short temperate and dry but are usually very cold for a few weeks Summers are typically long hot and relatively dry C653 Regarding the runoff produced in winter time in the case of the Passer basin we guess it is due to a combination of some snow melting and the abrupt change in the coefficient of runoff after the soil freezes e g Dall Amico et al 2011 However the confirmation of this will require further investigation and field work Q Model calibration is done twice and reasons for that are introduced quite abruptly in the conclusion section was just wondering what will happen if you calibrate the model only once and simulate the discharge for the entire year In this way the rationale for calibrating twice can be explained more easily and earlier in the text itself A The need for a double set of parameters in the original basin was generated by the presence of frozen soils in winter that changed the runoff generation dynamics This is one of the reasons why we changed the study basin and chose one with a more simple behavior In fact in the new simulations we indirectly answer your question by using one year for the calibration period and one year for the validation Q 5 When an analysis is done in section 3 am not sure if the readers are provided with the purpose and conclusions derived feel the same with the simulation results
9. Geosci Model Dev Discuss 4 C688 C675 2011 www geosci model dev discuss net 4 C638 201 1 Geoscientific i Sa Model Development Author s 2011 This work is distributed under Discussions the Creative Commons Attribute 3 0 License Interactive comment on The JGrass NewAge system for forecasting and managing the hydrological budgets at the basin scale the models of flow generation propagation and aggregation by G Formetta et al G Formetta et al formetta ing unitn it Received and published 9 September 2011 1 General answers to the reviewers questions We would firstly like to thank all the Reviewers for their interesting comments and sug gestions which will benefit the quality of our paper As can be verified in the following and in the newly submitted manuscript the paper has been greatly modified to satisfy the Reviewers requests However we have not added in depth descriptions of the C638 informatics infrastructure system and of the modelling of other budget components such as snow energy these are complex topic that require at least two separate papers in preparation for submission to GMD for their introduction and applications There have been four main changes e We decided to change the basin to which the model was applied because as suggested by reviewers in the case of the Passer river basin firstly the time series of measured data is too short and secondly snow is the main
10. Open File Report Section USGS Box 25425 Denver Co 80225 USGS Water Resources Investigations Report 83 4238 1983 207 p 54 fig 15 tab 51 ref 8 attach 1983 Mantilla R Gupta V et al Role of coupled flow dynamics and real network structures on Hortonian scaling of peak flows Journal of Hydrology 322 155 167 2006 Mesa O and Mifflin E On the relative role of hillslope and network geometry in hydrologic re sponse Scale Problems in Hydrology Eds V K Gupta I Rodriguez lturbe and E F Wood 1988 Montgomery D and Dietrich W A physically based model for the topographic control on shallow landsliding Water Resources Research 30 1153 1171 1994 Moore R The probability distributed principle and runoff production at point and basin scales Hydrol Sci J 30 165 1985 Naden P A routing model for continental scale hydrology Macroscale modelling of the hydro sphere 214 67 79 1993 Priestley C Turbulent transfer in the lower atmosphere University of Chicago Press Chicago Ill 1959 Priestley C and Taylor R On the assessment of surface heat flux and evaporation using large scale parameters Monthly weather review 100 81 92 1972 Rigon R D Odorico P and Bertoldi G The geomorphic structure of the runoff peak Hydrol ogy and Earth System Sciences Discussions 8 1031 1058 doi 10 5194 hessd 8 1031 2011 http www hydrol earth syst sci discuss net 8 1031 2011 2011 Rinaldo A
11. agement required however the need to estimate a whole set of hydrological quantities such as discharge evapotraspiration soil moisture bringing very soon to the implementation of more comprehensive modeling systems as the pioneering Stanford watershed model Crawford and Linsley 1966 or like the Sacramento model e g Burnash et al 1973 the PRMS model Leavesley et al 1983 They were usually based on the metaphor of intercommunicating compartments reservoirs each representing a process domain each one with its residence time C657 Q 945 20 21 As far as remember Beven 2001 does not talk only about lumped models He discusses all types of models Also check the reference style for distributed model references is missing A Correct we have changed the phrase to avoid misinterpretations We have also fixed the reference style Q 945 25 unnecessary is not a proper word here as it gives a differ ent meaning Perhaps something like without representing the full spatial variability would be better A Correction accepted the sentences have become In the first class the physics is modeled at grid pixels level using the fundamental laws of conservation of energy mass and momentum in the second the ruling equations are simplified in order to obtain some statistics of the hydrological budget without representing the full spatial variability of the processes Simplification
12. ality A This is the sentence on the paper The JGrass NewAge model Franceschi et al 2011 was conceived and structured to meet these demands to forecast not only if Coods but also of droughts to calculate the water balance at several points in the river network of a basin and to provide statistics revealing the internal C643 spatio temporal variability of some of the quantities analyzed To obtain this the model implements innovative informatics which is described in Antonello et al 2011 to allow modiinAcations of its parts and parameterizations without changing the whole and therefore supporting the comparison of different schemes of simpliifAAcation and of the parametrization of hydrological processes We agree with the reviewer However the sense of the phrase is given by without changing the whole and we in fact after having tried OpenMI now use an exist ing framework OMS3 The rationale of this choice will be explained in a companion paper but can be summarised as follows lower invasiveness of the framework in term of lines of codes Antonello et al 2011 intrinsic parallelization of components and the use of Java annotations to provide metadata about the model components In fact as explained in the paper the informatics of JGrass NewAGE use java an notations to automatically generate the graphic interfaces of the model components Furthermore this enhances the presence of a better community to s
13. be 1996 in which the topology and the geometry of the river network is more important than the details of the local dispersive dynamics e g Rinaldo et al 1991 Besides Rinaldo et al 1991 the statement is supported by the achievements of at least other two papers Rinaldo et al 1995 and D Odorico and Rigon 2003 where it was shown that the form of the instantaneous unit hydrograph as derived from the geomorphology is mainly affected by the geometry of flow paths both in hillslopes and channels Further studies on the same subject were made by Snell and Sivapalan 1994 Saco and Kumar 2004 Botter and Rinaldo 2003 Coincidentally Mesa and Mifflin 1988 and Naden 1993 used similar models to forecast flow peaks Furthermore it should not escape attention that during a flood event some simplifications of the dynamics occur as observed in the pioneering work by Leopold and Maddock in the 1950s Anyway we removed the sentence in the last version of the paper Q Line 4 6 Something is wrong with the sentence In addition flow ve locity at least don t understand what you try to say A The sentence is In addition Leopold and Maddock 1953 observed that the overall action of hydrological and geomorphological forces act in maintaining approxi mately constant the ow velocity We dropped the statement in the revised version C642 of the paper were we have stated that during floods the celerity
14. c conductivity depending on temperature ne glecting evapotranspiration and out ow from glaciers which were kept constant in C650 the present work since there was no way to assess their iniftCuence with measures We meant that neglecting some processes that are known to vary between summer and winter affected the remaining parameters values However the point about sat urated unsaturated conditions raised by the Reviewer was correct In the revised version we used a basin with no ice or snow and we introduced an explicit modeling of evapotranspiration The phrase above has been eliminated from the revised text C651 3 Answers to K Suprit Referee Q 1 In the title The models should be changed to a model or models authors may decide which one is more appropriate here Also it is easier for the readers if same order is maintained in headings in the text along with con sistent use of terminology in section 2 replace runoff production aggrega tion and routing with generation propagation routing and aggregation and accordingly modify the subsection heading 2 2 A Suggestion accepted The title will be The JGrass NewAge system for fore casting and managing the hydrological budgets at the basin scale models of inCow generation and propagation routing In the new version of the paper we decided to follow a more classical structure for the title of each stections in order
15. change was in fact the introduction of an automatic calibra tion algorithm for the model Specifically we applied a mono objective calibration procedure using the particle swarm optimization algorithm Kennedy and Eber hart 1995 Eberhart and Shi 2001 The calibration module is itself a reusable component of the JGrass NewAge System Furthermore all minor typing errors have been corrected and suggestions from re viewers accepted information about the features of the river basin and about the GIS udig JGrass have been added as requested by reviewers C640 2 Answers to P Krause Referee Q Abstract You mention that the system is able to estimate evapotranspira tion line 6 why don t you use this in your model application Same is true for snow modeling not mentioned at all in the application part Snow must be an important component in your catchment A For the revised text the simulations were repeated and evapotranspiration was computed with the Priestley Taylor model Given that snow modelling requires detailed validation and because snow is an important component of the water budget for many months of the year at the Passer river basin we decided to change the basin of study and use theLittle Washita river basin Oklahoma U S instead Therefore the paper can concentrate on the presentation and analysis of runoff generation and propagation routing components Q Page 945 Line 5 whole set of hydrological
16. component of the water budget in winter and spring time Therefore we moved to the Little Washita river basin located in Oklahoma U S where the USDA provides two year long time series of discharges and detailed hydro meteorological quantities Also the water balance of this basin is not influenced by snow We used the Priestley Taylor model Priestley 1959 Slatyer and Mcllroy 1961 Priestley and Taylor 1972 to assess evapotranspiration variability during the simulation years Priestley Taylor is just one of the models included in the JGrass NewAge System It is a very simple model and parameter parsimonious but it is adequate to the task required in the paper We decided to substitute for runoff production given by the Duffy model Duffy 1996b with that given by the Hymod model Moore 1985 and Boyle 2001 We were originally attracted by the conceptual simplicity of Duffy but we realized that this did not correspond to a simplicity of application since the parameter range provided in Duffy 1996b is not extensible to catchments with different soil types and soil hydraulic properties We could not therefore calibrate the runoff component model easily given that the parameters kept fixed in Duffy s original paper had to be calibrated and the model is not parsimonious On the contrary the Hymod runoff component has only five parameters which can be calibrated automatically more appropriately C639 e The fourth major
17. del in my opinion How model predictions relate to the observations should be discussed in more detail prior to the conclusions A In the revised version we have discussed the model prediction with more detail as requested Certainly a good fit is not a sufficient condition but a necessary one Q 14 p 955 Il 1 11 The conclusions regarding structural defects and signiinAcance of the differences in calibrated parameter values should be discussed in detail along with the results The fact that evapotranspiration and glacier outflow were kept constant should be mentioned before reporting the results along with the values they were held at A In the new simulations the evapotranspiration is not constant but computed using the Prestley Taylor method and the snow for the new basin Little Washita is not a relevant component of the water balance Q 15 p 955 Il 12 16 I am not sure that the authors clearly demon strate that these statistics of simulation are different or more reliable than other models or methods for doing so A We do not believe that our model is better than others but certainly a very good one Actually an unbiased comparison is probably impossible We have simply shown that the model fits the measures well However we do claim that in our framework a clean comparison can be made by changing the component under exam which is much more difficult within traditional monolithically implemented non co
18. dence times and how the mean and variance values are obtained Are they simply calibrated values If k and n are calibrated parameters as indicated below eq 10 then what is the significance of eq 11 and 12 C666 A We have eliminated this part from the new manuscript However the residence time was conceived as a distance divided by a velocity The distance was estimated by analyzing the structure of the paths in hillslopes following the drainage directions The velocity was calibrated Q 6 p 951 eq 13 How the terms of the right hand side are calculated and related to the terms in the preceding equations is not clearly shown It might be better to give the expression used for calculating Qi t in addition to its derivative shown in eq 13 Overall the underlying mathematics used for integration or scaling from unit area of hillslope to HRU and ultimately stream discharge is not entirely clear to me as seem to miss some calculation steps Is it possible to describe the links between stream discharge and processes at the unit area and HRU levels more clearly Consistent reuse of mathematical terms through the sequence of the equations would help in this regard A As suggested by the Reviewer we have better explained in the paper section Runoff propagation routing both the terms of the right hand side of the equation 13 and how they are related to the terms of the runoff generation component We have also presented the express
19. e impression that you are using nearly exactly the same semi distributed approach SWAT is using Maybe you like to refer that in your paper and discuss that a little bit A We know SWAT Our modeling effort can look similar in the objectives But SWAT uses a daily time scale while we would like to use an hourly time scale since the main characteristics of the response hydrograph are at a scale that is less than one day for most of basins up to some thousands of square kilometers Q Page 949 Eq 3 6 and the following explanations Please check very careful that you use the same indices and letters in text and equations Right now it is a little bit confusing A To better understand the meaning of each terms of the equations we suggest referal to Duffy 1996b and Duffy 1996a By all means this model is no longer used in the simulations presented in this paper Q Line 23 26 wonder how you derive d1 to d4 Do you calibrate them A We did not properly calibrate all of them but we did consider values different from those proposed in literature Duffy 1996a Duffy 1996a to take account of the soil properties of the Passer river basin These values were derived from the experience of one of the authors with the CUENCAS model If we should need to calibrate all of these parameters the task would easily become unfeasible For this reason in fact we decided to use the HyMod model as the runoff production component in the revised
20. he new revised paper for reasons explained above C656 Q iv Figures 7 and 8 are crucial for the paper Please be more careful in preparing them Check the captions dates on top of figure and in caption differ Your representation of dates x axis label seems to differ in both the panels Ideally it should be date and month Simulation not simulation data is enough in caption A We appreciate this suggestion and have modified the new figures accordingly Q 944 18 21 The model has been tested modeled discharge am not sure what is meant by scaling properties of discharge Also a statement regarding the results obtained or nature of simulations is missing A We have deleted the sentences regarding the scaling properties because it is not a key point in this paper Instead we have added information about the simulation period and the time step used in this paper The sentence has therefore became A two year simulation is presented at hourly time step in the Little Washita OK basin The model performances have been tested against measured discharges according to some classical indices of goodness of fit Q 945 5 10 in this context and the PRMS model Needs rewriting am confused with the statement about large modelling system Does it mean exhaustive data requirement or consideration of many processes A The sentence has been changed The water resource and river man
21. ich works for both periods Otherwise the single periods might be overcalibrated A We guessed that the parameters would be different since for instance in winter the soils of the higher parts of the Passer freeze and the rainfall runoff mechanism is completely changed Dall Amico et al 2011 In the new two year simulations we calibrated the model with one year s of data 2002 and validated it using the subsequent year s data 2003 Q Line 4 You mention structural model defects Firstly would call them structural model problems not defects Secondly please explain what problems you had in mind and maybe present solution how to solve them in the future A We agree that structural model problems were due to the fixed parameter in Duffy s model We have solved these problems by changing the runoff model used to a more parsimonious model where we could calibrate all the parameters Q Line 9 What do you mean by the dependence of hydraulic conductiv ity from temperature Do you think of frozen soils or what is it Is this really relevant as you neglect the dependence of hydraulic conductivity from soil water saturation maybe got that point wrong but then you should make it clearer in the paper A The sentence in the paper was Variation of parameters between summer and win ter which was necessary to obtain reasonably good results can be considered as a consequence of variation of hydrauli
22. ion used for calculating Qi t instead of its derivative as suggested by the Reviewer This is what we have written in the new section Runoff propagation routing The flow generation model along hillslopes delivers discharge to the channel network conceptualized in the model as a oriented tree graph For each link the continuity equation as presented in Mantilla et al 2006 is dSi t dt Qsen t gt Qiroli Qilt 1 2 5 o trib where the S t is storage in the link i th at time t H is the total number of network links Q t L3 T is the output discharge from i th link Qir L T is the flow of upstream links Qgen t L T is the discharge generated of the hillslope of the C667 link in question Under the hypothesis that the link has a rectangular x section so the width w does not change with time the channel storage and the discharge can be expressed as Silt li Wi d t 2 Qt vilt he wilt 4 d t 3 where v t L T 1 is the flow velocity w t L is the mean width of the link d t L is the mean channel depth and k L is the link length Combining the equations 2 and 3 gives S t in function of Q t finally using the Chezy equation v C R 79 4 where v L T is the mean velocity C L T is the Ch zy coefficient R L is the hydraulic radius and i is the bottom slope S can be expressed as 1 2 2 S t Q t
23. is from statistics is valid as there is only one peak in winter Can you think of a physical reason why the winter peaks are simulated better than summer peaks A We were aware that the winter simulations needed more work and have therefore switched to a basin where these interesting but challenging characteristics are not present Q 954 20 In the last paragraph some more detail is required It can be described as an advantage of the model Also add an and between link and therefore A Suggestions accepted the new last paragraph presents more details and we added the and between link and therefore Q 954 25 am not sure if this conclusion follows from the text Some sort of explanation is required A In the new revised text we have changed the conclusions C662 Q 955 5 Vague statement about rainfall measurement errors Needs de tailed statement The point made in text can be repeated here A We have removed these comments Q 955 15 any link end Not clear A We have removed these considerations from the paper C663 4 Answers to S Samanta Referee Q 1 p 946 Il 12 19 many modern hydrological models are able to simulate stream flow as a time series i e not restricted to modeling imCoods or droughts and are able to provide estimates at various locations Many process oriented hydrologic models also employ component structures due to its various advan tage
24. ised version of the paper Q 10 p 952 Il 21 25 Include an explanation of how the area perimeter relationship results of the linear regression the relationship P A 489 is not linear the mean and the variance mentioned here are used in the context of the model equations described before The mean and variance mentioned in eq 11 and 12 relate to distributions of residence times If these two sets of mean and variance values are considered related please include a justification for that A We have removed this part because it is not relevant in this application The perimeter has a non linear relation with area because it is a fractal e g Rodriguez C669 Iturbe and Rinaldo 2001 Q 11 p 953 Il 5 15 Please mention which parameters were cali brated and which were kept constant along with the values used for the summer and winter simulations so that their signiinAcance may be clearly understood by the readers As the model structure is process oriented the calibrated parameter values and their differences between seasons may be interesting and informative about the system In addition many modern automatic calibration methods provide some advantages e g uncertainty estimates over manual calibration methods Is there any specific reason behind choosing a manual method A No there was not a specific reason In fact in the new simulations we have used an automatic calibration system as specified in the gene
25. izzi S Gruber S and Rigon R A robust and energy conserving model of freezing variably saturated soil The Cryosphere 5 469 484 doi 10 5194 tc 5 469 2011 http www the cryosphere net 5 469 2011 2011 Dietrich W Wilson C Montgomery D McKean J and Bauer R Erosion thresholds and land surface morphology Geology 20 675 1992 D Odorico P and Rigon R Hillslope and channel contributions to the hydrologic response Water resources research 39 1113 2003 Duffy C A two state integral balance model for soil moisture and groundwater dynamics in complex terrain Water resources research 32 2421 2434 1996a Duffy C J A Two State Integral Balance Model for Soil Moisture and Groundwater Dy namics in Complex Terrain Water Resour Res 32 2421 2434 http dx doi org 10 1029 96WR01049 1996b Eberhart R and Shi Y Particle swarm optimization developments applications and re sources in Proceedings of the 2001 congress on evolutionary computation vol 1 pp 81 86 Piscataway NJ USA IEEE 2001 C673 Franceschi S A A G F and R R NeAge the user manual In preparation 2011 Kennedy J and Eberhart R Particle swarm optimization in Neural Networks 1995 Pro ceedings IEEE International Conference on vol 4 pp 1942 1948 IEEE 1995 Leavesley G Lichty R Troutman B and Saindon L Precipitation runoff modeling system User s manual Available from Books and
26. lated analyses network related measures and geomorphological classifications tools for hillslope analysis and for performing DEM statistics Finally the jgrasstools contain other modeling efforts such as SHALSTAB Dietrich et al 1992 and Montgomery and Dietrich 1994 Peakflow e g Rigon et al 2011 and others Q 3 get an impression that summer rainfall is 2 3 times more than win ter see the different ordinate scales in Fig 7 0 5 mm h and 8 0 15 mm h yet the highest discharge peak is seen in winter It may be interesting for readers who are not familiar with the region to have some background of the physical and climatic setting of the basin along with observed intra annual variability in precipitation etc A We thank the reviewer for the suggestion In the re drafted paper we have given some background of the physical and climatic setting of the Little Washita basin and discuss in more detail the data avaialble Added The Little Washita river basin 611 square kilometers is located in south west Oklahoma between Chickasha and Lawton The elevation ranges between about 500 meters and about 300 meters Allen and Naney 1991 The bedrock exposed in the watershed consists of Permian age sedimentary rocks and soil textures range from fine sand to silty loam Allen and Naney 1991 The climate in the basin can be characterized as moist and sub humid with a long term spatially average annual precipitation of
27. mponent based models C671 Q 16 The captions for the figures should be more descriptive and un derstandable on their own as far as possible A We accept the correction and have modified the layout of the figures C672 References Ahmed S and Marsily G D Comparison of geostatistical methods for estimating transmissiv ity using data on transmissivity and specific capacity Water resources research 23 1987 Allen P and Naney J Hydrology of the Little Washita River watershed Oklahoma data and analyses ARS US Department of Agriculture Agricultural Research Service USA 1991 Antonello A Franceschi S Formetta G and Rigon R The JGrass NewAge System for Forecasting and Managing the Hydrological Budgets at the Basin Scale the informatic struc ture 2011 Botter G and Rinaldo A Scale effect on geomorphologic and kinematic dispersion Water resources research 39 1286 2003 Boyle D P Multicriteria calibration of hydrological model Ph D dissertation Dep of Hydrol and Water Resour Univ of Ariz Tucson 2001 Burnash R Ferral R and McGuire R A generalized streamflow simulation systemaAT Con ceptual modeling for digital computers report Joint Fed and State River Forecast Cent US Natl Weather Serv Calif State Dept of Water Resour Sacramento Calif 1973 Crawford N and Linsley R Digital simulation in hydrology Standford watershed model 4 1966 Dall Amico M Endr
28. now and ice processes could they be relevant for the underestimation as well A The original spatial estimates of precipitation were made using a kriging with elevation drift Anmed and Marsily 1987 one of the more used algorithms in mountain application However in the new simulations on the Little Washita river basin we have used an ordinary kriging to interpolate the precipitation data because the range of elevation of this basin is less than 200 m While the snow melting and the ice certainly contributed to the underestimation of discharges in the Passer Valley the problem is no longer present in the case of Little Washita Q Page 954 line 20 21 Indeed you show that the model is able to show the hydrographs But from fig 9 have the impression that the model is not able to calculate them correctly All of them look awkward with that long recession dominating the entire hydrographs Looks like a linear storage which starts with a large content and which continuously releases water without being replenished Some discussion on that would be interesting and absolutely needed A These problems are not anymore present in the new version of the paper C649 Q Page 955 line 3 You should show us the parameters and you should try to explain which parameter is different and why As you have stated earlier in the paper that the parameters have a physical meaning it is feasible to do so Are you able to define one parameter set wh
29. nsions in eqs 3 6 I guess As should be dimensionless Please give the units of Ds What about unit of A in eq 13 A To better understand the meaning of each term of the equations we refer you to Duffy 1996b and Duffy 1996a By all means the Duffy model is no longer used for the simulations presented in the revised paper for the reasons explained in answering the first Reviewer prof Krause Q i Figure 1 is nice However some improvement is required Bigger arrows and more details embedded in the schematic will be appreciated What does the right most arrow depict A The right most arrow had the same meaning as fz outgoing boundary flux on Sp However the figure is not present in the revised paper Q ii Figures 2 and 5 can be combined together 2 or 5 can be used in inset The intermediate positions of figure 9 can also be marked here Also in Figure 2 please check In caption outlet of Passer river basin is at Saltusio but in the text it is at Bojen A We set the entire model of the Adige river basin with outlet to Bozen as we wrote in the text but the sub basin where we tested the model calibrated and validated is in the sub basin with outlet to Saltusio this is the Passer river basin The motivation of this choice was the fact that the Passer river basin discharge is not affected by anthropogenic modifications such as hydropower power plants artificial reservoirs However the Passer basin is not used in t
30. nsistent terminology A We accept the comment We use model or system through all the text Q 952 19 gemorphological geo A We have removed this part because it was not useful for the model it served only to illustrate the geomorphological features of the basin Q 952 20 If not given earlier some reference for JGrass will be appre ciated here A We have added some information and citations about Jgrass at the beginning of section 2 Q 952 21 The relationship area perimeter insert between A We have removed this part because it is not essential Q 952 22 In Fig 4 it is It is A We have removed this part C661 Q 953 20 Please be more clear how you arrive at this conclusion where volume of discharge is considerable Illy greater than precipitation From the Figs 7 and 8 can understand it for the winter peak but for summer am not sure One way to see it will be to compare areal precipitation with discharge A The new simulations on the Little Washita show different dynamics Q 954 8 some examples of state of art models will be helpful here A The sections in which results are presented and discussed were modified in order to satisfy the reviewers requests As far this particular sentences it was removed from the paper We cited milestones in modeling throughout the whole new texts Q 954 13 am not sure inferring th
31. of flood waves tends to be constant in space as supported by the experiments of Leopold and Maddock and others and by some theoretical speculations Rodriguez lturbe et al 1992 This statement allows us to conceive a simpler modellization of the flood wave keeping this quantity the velocity of flow in channels fixed Q Line 6 either prediction is or predictions are A Corretion accepted Q Line 7 11 To this end this sentence is too exclusive in my opin ion Of course RS provides very helpful information for spatial distributed modeling but it is not a must must as you indicate in your text A The sentence was To this end it is necessary to make use of detailed information on topography as derived from modern LIDAR or SAR sensors and a large variety of remote sensed information which provide new tools for representation of the physics of Cow transport along the channels of the river network and processes into the hillslopes The sentence was removed in the last version of the paper Q Line 17 modification of model parameterisation is not an innovative informatics as you indicate in your text This possible and necessary with for nearly every model that needs to be transferred and calibrated The mentioned modification of its parts is more innovative but there are several modeling framework systems available e g JAMS OMS TIME MMS etc which provide such function
32. r velocity to vary at each stage and is not therefore treated as a constant as is common in flow peak models By all means the temporal scale at which the model works is hourly or sub hourly Q 946 25 Any HRU instead this statement seems vague Also His are important here so 2 3 lines can be devoted to explain how they are delin eated A Suggestion accepted An explanation of the process of hillslope delineation was added in the revised text Q 947 23 Title can highlight ADIGE Jgrass A J Grass NewAge is the complete system but the object of this paper is to present the Adige component only which includes the processes of runoff production aggregation and routing Q 947 25 Where L means a length such as mm something like L is length mm C659 A L means a generic length unit it can be mm or m according to the other unit of the terms appearing in the equations We have made this more explicit in the new text Q 948 5 Differently form most of the models Is the model used different from above mentioned models only or some other models Are there any similar models A Because we changed the runoff generation component this sentence was removed in the new version of the paper Q 948 5 coupled and generate runoff gt meaning is not clear A The sentence is The equations of the two reservoirs are coupled and generate runoff while the r
33. ral comment section The calibrated parameters of the new runoff production component are now more clearly specified Q 12 Deficiencies in data p 953 Il 19 24 and model structure p 955 ll 4 5 are cited as reasons for some systematic errors Such errors are usually more informative for making modeling improvements compared to simple goodness of fit measures would suggest showing the systematic deviations more clearly perhaps using a residual plot and including a more thorough discussion of such deviations A We agree with the observations and comments of the Reviewer The data quality was another problem of the old watershed in the new application on the Little Washita river basin as you can see in the simulations the quality of the data is better For the new simulations we have adopted the suggestions of plotting of residual errors However this plot needed to be commented since small time shifts between the measured and simulated discharges cause in some events large differences which could be not very relevant C670 Q 13 p 954 Il 23 25 The idea of component based model structure is not new and its use for this model is not really demonstrated in this paper Also the details of the informatics structure are not described in the paper Therefore these may not be considered as conclusions from this study itself Moreover good fit to a specific data set may not be considered as sufficient validation of a mo
34. rical models of the Earth System and its components In fact the use of the new informatics allows for seamless inclusion of the calibration algorithm based on particle swarming without any modification to the rest of the code In fact the model with calibration and without calibration are two different assemblies of the same basic components but using different Groovy scripting Q 7 There should be a separate subsection or at least a short paragraph about datasets terrain precipitation etc used It will be also good to discuss the data requirements of the modeling system A We accept the suggestion About the datasets used we have added to the basin general information with more information in the section Application This part will also clarify the data requirements Q 8 Authors claim that the modeling system is built upon flexible com ponents This is not coming out clearly The modelling system is built on Jgrass an open source and free GIS This point should also be highlighted A We accept the suggestion and we have added this information in section 2 as this modification also was called for in answer to previous comments Q Figures Figures need considerable improvement Expand and check C655 all captions carefully Figures 7 8 9 are hard to comprehend especially the axes labels A We accept the suggestions and we have done our best to improve figure readibility Q 9 Check subscripts and dime
35. s and also use the concept of hydrological response unit Therefore these capabilities of the JGrass NewAge model do not appear to be new or different The authors should clearly convey the specific advantages of their model in com parison to models that have similar capabilities instead of why these capabilities are important in the context of hydrological modeling in general as currently done in the preceding paragraphs Moreover the ability to provide statistics revealing the internal spatio temporal variability of some of the quantities ana lyzed is very imprecise while the details of the innovative informatics are not readily accessible as the Antonello et al 2011 is indicated to be in preparation stage These should be explained in greater detail A Dr Samanta is certainly right to many respects and we have tried to answer to his comments and to avoid generic statements in the new version of the paper please see also answers to comments to the other reviewers One fact is that all the above characteristics he mentions are usually not possessed by a single model but can be found spread in different models As far as our effort is concerned we have tried not to focus on a single model but on the possibility to change part of the model at run time not at compiling time using the OMS3 infrastructure about which appropriate doc umentation can be found at http Awww javaforge com project oms This framework in turn wa
36. s not developed by us but we have add to it the ability to be integrated without problems in a full featured GIS system udig jgrass of which information and the source code can be found at http udig refraction net whereas the models we are talking about are available at http code google com p jgrasstools In fact the new Spatial toolbox which is freely available with the new version is largely an outcome C664 of the research presented in this paper Therefore the JGrass NewAge users have in principle the possibility to mix model components and create their own run time model and prepare and visualize the results inside the GIS system This paper how ever does not talk about the core informatics that made it possible that actually even if with scarce but increasing documentation is freely available and open to inspection by researchers but has the goal to demonstrate that the strategy adopted is effective Therefore we assembled submodels into the JGrass NewAge and with the paper we aim to demonstrate that this is a very good way to reproduce the discharges in a catch ment To achieve these results we actually used the ability of joining components by adding to the model tools for the spatial distribution of rainfall inputs tools for calibra tion in this case the particle swarming algorithm and added in the revised version of the paper the model of Priestley Taylor to estimate evapotranspiration Discussions were made in
37. sometimes were derived by solid arguments and in this context a solid paradigm is offered by the theory of the geomorphological unit hydrograph e g Rinaldo et al 1991 Rigon et al 2011 or by heuristic subjective arguments Q 945 26 last para It may read better if this paragraph is linked to the preceding paragraph building from whether GIUH is lumped or distributed A This paragraph is no longer part of the article because we have replaced the runoff generation component with the Hymod model Q 946 7 discharge at intranet location got the meaning but won der if a more suitable replacement for intranet can be found A We intented it is a word that gives the idea of internal points of the network to avoid confusion we have used this periphrasis in place of the word intranet in the revised text C658 Q 946 10 physics not required am also not sure how remote sensing provides new tools it should be new data A The sentence was removed in the new version of the paper Q 946 12 floods and draughts Is it necessary to bring it here without mentioning the speciinAc spatial temporal scales at which Jgrass NewAge works A We contend that it is necessary because it is not only a model capable of capturing flood events but it also captures droughts This is also thanks to the propagation module of the generated runoff which for instance allows for wate
38. terms of water balance etc and linking eq 1 and 2 concepts to these equations would be useful for understanding the physical basis of the model A These concepts are presented in the papers Duffy 1996b and Duffy 1996a Anyway as reported in the paper The model is based on integration of the continuity equation over a hillslope control volume This is made up of the saturated and unsaturated soil storage with the water table serving as a moving boundary between these two storage volumes The Reynolds transport theorem is used to relate local continuum equations for moisture to the system storages and uxes and the diver gence theorem relates integrals of the spatial derivatives to integrals of the surface inCuxes see Duffy 1996b for details The Reviewer should kindly notice that even we answer to the questions he raised on the Duffy s model we have changed it in the final revision of the paper Q 3 p 949 Il 10 22 The superscript 0 of S is not defined Are S1 and S2 the same as S01 and S02 A No they are not the same S1 and S2 represent the volume of water in unsatu rated and saturated soil fractions per unit area of hillslope on the contrary S01 and S02 are the residual storage volumes in unsaturated and saturated soil fractions per unit area of hillslope Q 5 p 951 eq 10 11 12 Please analytically explain or justify the use of this simplification based on the use of mean and variance of resi
39. this paper Spatial toolbox aware OMS3 components are able to deal with spatial coverages and features according to the standards of the Open Spatial Consortium OGC as implemented by the Geotools http Awww geotools org and work seamlessly inside the GIS within a user interface that is automatically created from the metadata present in any component s source code appropriately programmed Therefore the JGrass NewAGE users have the possibility to mix compatible model components and create their own run time model and prepare and visualize the re sults inside the GIS system This paper however does not talk about the core informatics that made it possible but has the goal to demonstrate that the strategy adopted is effective Therefore we assembled some submodels into the JGrass NewAGE and with the paper we aim to demonstrate that this assembly is able to reproduce at best the discharges in a catchment The model we structured out of the JGrass NewAGE components was addressed to forecast the whole hydrographs at hourly time scale to calculate the water budget at several points inside the river network of a basin and to give an estimate of the evapotranspiration Q The verb confounded might not really indicate what you had in mind Think of a more proper word C645 A The sentence was this concept could be confounded The sentence is changed in this concept could be confused Q From your description get th
40. to take in account of the reviewes suggestions Abstract 1 Introduction 2 Methods 2 1 Basin delineation 2 2 Rainfall and other Input data 2 3 Runoff generation 2 4 Flow Routing 2 5 Evapotranspiration 2 6 Parameters estimation 2 7 Verification 3 An Application to Little Washita OK river basin 3 1 Results 4 Conclusions Q 2 This paper assumes that readers have a prior knowledge of Jgrass NewAge in frastructure from literature Since the paper and manual referred to in the manuscript are in preparation this is not essentially true So a little more detail in this regard is required A Even though the Jgrass NewAge system itself is not the object of this paper we have added some information about it at the beginning of the section 2 JGrass NewAge is implemented within the udig GIS system using its new spatial toolbox which is an achievement partially generated by this project based on OMS3 that permits the execution of any of the jgrasstools http code google com p jgrasstools with a proper C652 graphical user interface as part of the GIS All the jgrasstools modules are open source distributed with source code under the GPL v 3 0 license and programmed in Java but modules written in C C or FORTRAN could also be included Besides the components written specifically for JGrass NewAGE the jgrasstools include basic and less trivial routines for digital elevation model DEM manipulation basin re
41. unoff routing itself is described by a simple distribution of residence times The term coupled refers to the method of solving the equations mathemati cally i e they are solved as a system of equations Once the system of equations has been solved runoff is generated Q 950 1 6 First line of this paragraph is quite abrupt think a re organisation or rewrit ing will make better sense Actually the second sentence should start this section A As suggested we have started with the second sentence Q 950 10 In this calculations Darcy s law average according con sider rewriting A Darcy s law is not used in the new draft of the paper Q 951 17 Eq 13 Please check the dimensions Is it A area or frac tion of area A It is the saturated fraction area a fraction of the total area dimensionless Q 952 5 Menabde and Sivapalan 2001 missing from reference C660 A Added Q 952 Section 3 heading passer to Passer A We accept the correction Q 952 11 and shown in Fig delete It is a good practice to avoid as much as possible usages such as shown in Fig or as shown or given in Table etc Figures and tables should be mentioned in parenthesis along with concerned statements A For the new figures we have taken into account your suggestion Q 952 18 Jgrass NewAge infrastructure model or infrastructure or stat Use co
42. upport common developments The fact that we did not create OMS3 does not means that it is not innovative However we believe that we have added value to it making available OGC standards and a the Spatial Tooolbox now a standard component of the udig GIS http udig refractions net to OMS3 modelers They are we believe an innovation within the innovation We rephrased as follows The JGrass NewAge was conceived as an infrastructure able to offer a graphical in terface to models without extra programming burden Antonello et al 2011 in prepa ration and to reduce the gap between the production of new research and its deploy ment to stakeholders For achieving this result after testing alternatives the OMS3 infrastructure http oms javaforge com was chosen having found in its concepts of programming by components which allows to test different modeling strategies by changing parts of C644 a model at run time not at compiling time a strategy either useful to research and to tailor modeling solutions to the requirement of a particular use case To Jgrass NewAGE besides the OMS3 framework capabilities we added the inte gration in a full featured geographic information system GIS system udig jgrass http udig refraction net and http Awww jgrasstools net In fact the new udig s Spa tial toolbox which is freely available with the new version of the GIS is largely an outcome of the research presented in
43. where we used a two year long time series of discharges with an hourly time step Q Line 6 Use sensitive or dominant instead of influential A Done Q Line 9 Why didn t you calculate ET a constant ET over a period of several months is not very likely Here the question arises Why do you get such a good model performance If assume that ET is something around 50 just a guess please correct me or more of the precipitation amount in summer this should be relevant A Prof Krause is correct In our new experiment we have decided to compute evapotranspiration using the Priestley Taylor model In practice this is another OMS3 component also implemented in JGrass NewAge or which is the same in the jgrasstools A more complete validation of this and other models remains however the argument of other incoming papers C648 Q Line 15 ff Something went wrong in the description and the figure leg end Now am confused what is observed what is simulated runoff A We have redrafted the figure to make it clearer In all the new figures the gray dots represent the measured discharge and the solid black line represents the simulated one Q Line 19 ff cannot really see the underestimation you mention In the latter part of this paragraph that the underestimation of discharge is caused by an underestimation of rainfall How do you make the spatial estimates don t you account for changes in elevation And what about s
Download Pdf Manuals
Related Search