• Archibald, J. A., and M. T. Walter, 2014: Do energy‐based PET models require more input data than temperature‐based models?—An evaluation at four humid FluxNet sites. J. Amer. Water Resour. Assoc., 50, 497508, doi:10.1111/jawr.12137.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Balling, R. C., Jr., and G. B. Goodrich, 2011: Spatial analysis of variations in precipitation intensity in the USA. Theor. Appl. Climatol., 104, 415421, doi:10.1007/s00704-010-0353-0.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Bell, V., A. Kay, H. Davies, and R. Jones, 2016: An assessment of the possible impacts of climate change on snow and peak river flows across Britain. Climatic Change, 136, 539553, doi:10.1007/s10584-016-1637-x.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Blöschl, G., and Coauthors, 2007: At what scales do climate variability and land cover change impact on flooding and low flows? Hydrol. Processes, 21, 12411247, doi:10.1002/hyp.6669.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Blöschl, G., and Coauthors, 2015: Increasing river floods: Fiction or reality? Wiley Interdiscip. Rev.: Water, 2, 329344, doi:10.1002/wat2.1079.

  • Bouwer, L. M., 2011: Have disaster losses increased due to anthropogenic climate change? Bull. Amer. Meteor. Soc., 92, 3946, doi:10.1175/2010BAMS3092.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Brigode, P., and Coauthors, 2014: Sensitivity analysis of SCHADEX extreme flood estimations to observed hydrometeorological variability. Water Resour. Res., 50, 353370, doi:10.1002/2013WR013687.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Butler, D. R., 1989: The failure of beaver dams and resulting outburst flooding: A geomorphic hazard of the southeastern Piedmont. Geogr. Bull., 31, 2938.

    • Search Google Scholar
    • Export Citation
  • Ceppi, A., G. Ravazzani, A. Salandin, D. Rabuffetti, A. Montani, E. Borgonovo, and M. Mancini, 2013: Effects of temperature on flood forecasting: Analysis of an operative case study in Alpine basins. Nat. Hazards Earth Syst. Sci., 13, 10511062, doi:10.5194/nhess-13-1051-2013.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Condon, L. E., S. Gangopadhyay, and T. Pruitt, 2015: Climate change and non-stationary flood risk for the upper Truckee River basin. Hydrol. Earth Syst. Sci., 19, 159175, doi:10.5194/hess-19-159-2015.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Dahlke, H. E., Z. M. Easton, D. R. Fuka, S. W. Lyon, and T. S. Steenhuis, 2009: Modelling variable source area dynamics in a CEAP watershed. Ecohydrology, 2, 337349, doi:10.1002/eco.58.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • DeGaetano, A. T., 2009: Time-dependent changes in extreme-precipitation return-period amounts in the continental United States. J. Appl. Meteor. Climatol., 48, 20862099, doi:10.1175/2009JAMC2179.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Diffenbaugh, N. S., J. S. Pal, R. J. Trapp, and F. Giorgi, 2005: Fine-scale processes regulate the response of extreme events to global climate change. Proc. Natl. Acad. Sci. USA, 102, 15 77415 778, doi:10.1073/pnas.0506042102.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Easton, Z. M., P. Gérard-Marchant, M. T. Walter, A. M. Petrovic, and T. S. Steenhuis, 2007: Hydrologic assessment of an urban variable source watershed in the northeast United States. Water Resour. Res., 43, W03413, doi:10.1029/2006WR005076.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Elbialy, S., A. Mahmoud, B. Pradhan, and M. Buchroithner, 2014: Application of spaceborne synthetic aperture radar data for extraction of soil moisture and its use in hydrological modelling at Gottleuba catchment, Saxony, Germany. J. Flood Risk Manage., 7, 159175, doi:10.1111/jfr3.12037.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Freudiger, D., I. Kohn, K. Stahl, and M. Weiler, 2014: Large scale analysis of changing frequencies of rain-on-snow events and their impact on floods. Hydrol. Earth Syst. Sci., 18, 26952709, doi:10.5194/hess-18-2695-2014.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Frumhoff, P. C., J. J. McCarthy, J. M. Melillo, S. C. Moser, and D. J. Wuebbles, 2007: Confronting climate change in the U.S. Northeast: Science, impacts, and solutions. Northeast Climate Impacts Assessment Rep., Union of Concerned Scientists, 146 pp. [Available online at http://www.ucsusa.org/sites/default/files/legacy/assets/documents/global_warming/pdf/confronting-climate-change-in-the-u-s-northeast.pdf.]

  • Fuka, D. R., M. T. Walter, J. A. Archibald, T. S. Steenhuis, and Z. M. Easton, 2013: EcoHydRology: A community modeling foundation for eco-hydrology, version 4.9. R package. [Available online at https://cran.r-project.org/web/packages/EcoHydRology/index.html.]

  • Genest, C., and A. C. Favre, 2007: Everything you always wanted to know about copula modeling but were afraid to ask. J. Hydrol. Eng., 12, 347368, doi:10.1061/(ASCE)1084-0699(2007)12:4(347).

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Gersonius, B., R. Ashley, A. Pathirana, and C. Zevenbergen, 2013: Climate change uncertainty: Building flexibility into water and flood risk infrastructure. Climatic Change, 116, 411423, doi:10.1007/s10584-012-0494-5.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Gilroy, K. L., and R. H. McCuen, 2012: A nonstationary flood frequency analysis method to adjust for future climate change and urbanization. J. Hydrol., 414–415, 4048, doi:10.1016/j.jhydrol.2011.10.009.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Grusson, Y., X. Sun, S. Gascoin, S. Sauvage, S. Raghavan, F. Anctil, and J. M. Sáchez-Pérez, 2015: Assessing the capability of the SWAT model to simulate snow, snow melt and streamflow dynamics over an alpine watershed. J. Hydrol., 531, 574588, doi:10.1016/j.jhydrol.2015.10.070.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Haberlandt, U., and I. Radtke, 2014: Hydrological model calibration for derived flood frequency analysis using stochastic rainfall and probability distributions of peak flows. Hydrol. Earth Syst. Sci., 18, 353365, doi:10.5194/hess-18-353-2014.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Harrison, B., and R. Bales, 2015: Skill assessment of water supply outlooks in the Colorado River basin. Hydrology, 2, 112131, doi:10.3390/hydrology2030112.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Hayhoe, K., and Coauthors, 2007: Past and future changes in climate and hydrological indicators in the US Northeast. Climate Dyn., 28, 381407, doi:10.1007/s00382-006-0187-8.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Hayhoe, K., and Coauthors, 2008: Regional climate change projections for the Northeast USA. Mitigation Adapt. Strategies Global Change, 13, 425436, doi:10.1007/s11027-007-9133-2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Hirabayashi, Y., R. Mahendran, S. Koirala, L. Konoshima, D. Yamazaki, S. Watanabe, H. Kim, and S. Kanae, 2013: Global flood risk under climate change. Nat. Climate Change, 3, 816821, doi:10.1038/nclimate1911.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Hirsch, R. M., and K. R. Ryberg, 2012: Has the magnitude of floods across the USA changed with global CO2 levels? Hydrol. Sci. J., 57, 19, doi:10.1080/02626667.2011.621895.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Jabareen, Y., 2013: Planning the resilient city: Concepts and strategies for coping with climate change and environmental risk. Cities, 31, 220229, doi:10.1016/j.cities.2012.05.004.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Jörg-Hess, S., N. Griessinger, and M. Zappa, 2015: Probabilistic forecasts of snow water equivalent and runoff in mountainous areas. J. Hydrometeor., 16, 21692186, doi:10.1175/JHM-D-14-0193.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Knighton, J., and M. Walter, 2016: Critical rainfall statistics for predicting watershed flood responses: Rethinking the design storm concept. Hydrol. Processes, 30, 37883803, doi:10.1002/hyp.10888.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Koplin, N., B. Schadler, D. Viviroli, and R. Weingartner, 2014: Seasonality and magnitude of floods in Switzerland under future climate change. Hydrol. Processes, 28, 25672578, doi:10.1002/hyp.9757.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Kundzewicz, Z. W., and Coauthors, 2014: Flood risk and climate change: Global and regional perspectives. Hydrol. Sci. J., 59, 128, doi:10.1080/02626667.2013.857411.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Kunkel, K. E., and Coauthors, 2013: Monitoring and understanding trends in extreme storms: State of knowledge. Bull. Amer. Meteor. Soc., 94, 499514, doi:10.1175/BAMS-D-11-00262.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Lawrence, D., E. Paquet, J. Gailhard, and A. K. Fleig, 2014: Stochastic semi-continuous simulation for extreme flood estimation in catchments with combined rainfall–snowmelt flood regimes. Nat. Hazards Earth Syst. Sci., 14, 12831298, doi:10.5194/nhess-14-1283-2014.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Li, J., P. Feng, and F. Chen, 2014: Effects of land use change on flood characteristics in mountainous area of Daqinghe watershed, China. Nat. Hazards, 70, 593607, doi:10.1007/s11069-013-0830-8.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Lo, M.-H., and J. S. Famiglietti, 2010: Effect of water table dynamics on land surface hydrologic memory. J. Geophys. Res., 115, D22118, doi:10.1029/2010JD014191.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Mahanama, S., B. Livneh, R. Koster, D. Lettenmaier, and R. Reichle, 2012: Soil moisture, snow, and seasonal streamflow forecasts in the United States. J. Hydrometeor., 13, 189203, doi:10.1175/JHM-D-11-046.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Mallakpour, I., and G. Villarini, 2015: The changing nature of flooding across the central United States. Nat. Climate Change, 5, 250254, doi:10.1038/nclimate2516.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Massari, C., L. Brocca, S. Barbetta, C. Papathanasiou, M. Mimikou, and T. Moramarco, 2014a: Using globally available soil moisture indicators for flood modelling in Mediterranean catchments. Hydrol. Earth Syst. Sci., 18, 839853, doi:10.5194/hess-18-839-2014.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Massari, C., L. Brocca, T. Moramarco, Y. Tramblay, and J. F. D. Lescot, 2014b: Potential of soil moisture observations in flood modelling: Estimating initial conditions and correcting rainfall. Adv. Water Resour., 74, 4453, doi:10.1016/j.advwatres.2014.08.004.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Massari, C., A. Tarpanelli, L. Brocca, and T. Moramarco, 2015: Assimilating satellite soil moisture into rainfall–runoff modelling: Towards a systematic study. Geophysical Research Abstracts, Vol. 17, Abstract EGU2015-1780. [Available online at http://meetingorganizer.copernicus.org/EGU2015/EGU2015-1780.pdf.]

  • Moriasi, D. N., J. G. Arnold, M. W. Van Liew, R. L. Bingner, R. D. Harmel, and T. L. Veith, 2007: Model evaluation guidelines for systematic quantification of accuracy in watershed simulations. Trans. ASABE, 50, 885900, doi:10.13031/2013.23153.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Nash, J. E., and J. V. Sutcliffe, 1970: River flow forecasting through conceptual models part I—A discussion of principles. J. Hydrol., 10, 282290, doi:10.1016/0022-1694(70)90255-6.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Nied, M., Y. Hundecha, and B. Merz, 2013: Flood-initiating catchment conditions: A spatio-temporal analysis of large-scale soil moisture patterns in the Elbe River basin. Hydrol. Earth Syst. Sci., 17, 14011414, doi:10.5194/hess-17-1401-2013.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • NOAA/HDSC, 2016: Precipitation Frequency Data Server (PFDS). Accessed 28 December 2016. [Available online at http://hdsc.nws.noaa.gov/hdsc/pfds/index.html.]

  • NOAA/NCEI, 2015: Quality controlled local climatological data. Accessed 28 December 2016. [Available online at http://www.ncdc.noaa.gov/data-access/land-based-station-data/land-based-datasets.]

  • Norbiato, D., M. Borga, S. Degli Esposti, E. Gaume, and S. Anquetin, 2008: Flash flood warning based on rainfall thresholds and soil moisture conditions: An assessment for gauged and ungauged basins. J. Hydrol., 362, 274290, doi:10.1016/j.jhydrol.2008.08.023.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • NRCC, 2015: NRCC hourly precipitation database. Accessed 28 December 2016. [Available online at http://www.nrcc.cornell.edu/services/access/access.html.]

  • NRCS, 2015: Web Soil Survey. Accessed 11 November 2016. [Available online at http://websoilsurvey.sc.egov.usda.gov/App/HomePage.htm.]

  • Obeysekera, J., and J. D. Salas, 2014: Quantifying the uncertainty of design floods under nonstationary conditions. J. Hydrol. Eng., 19, 14381446, doi:10.1061/(ASCE)HE.1943-5584.0000931.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Palecki, M. A., J. R. Angel, and S. E. Hollinger, 2005: Storm precipitation in the United States. Part I: Meteorological characteristics. J. Appl. Meteor., 44, 933946, doi:10.1175/JAM2243.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Paquet, E., F. Garavaglia, R. Garçon, and J. Gailhard, 2013: The SCHADEX method: A semi-continuous rainfall–runoff simulation for extreme flood estimation. J. Hydrol., 495, 2337, doi:10.1016/j.jhydrol.2013.04.045.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Paschalis, A., S. Fatichi, P. Molnar, S. Rimkus, and P. Burlando, 2014: On the effects of small scale space–time variability of rainfall on basin flood response. J. Hydrol., 514, 313327, doi:10.1016/j.jhydrol.2014.04.014.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Perju, E. R., D. Balin, S. Lane, and L. Zaharia, 2013: Changing flood magnitude and frequency in snow-melt dominated catchments: The case of the Bucegi Mountains, in the Romanian Carpathian region. Geophysical Research Abstracts, Vol. 15, Abstract EGU2013-10110. [Available online at http://meetingorganizer.copernicus.org/EGU2013/EGU2013-10110.pdf.]

  • Ravazzani, G., M. Mancini, I. Giudici, and P. Amadio, 2007: Effects of soil moisture parameterization on a real-time flood forecasting system based on rainfall thresholds. IAHS Publ., 313, 407–416.

  • Rossman, L. A., 2010: Storm water management model user’s manual, version 5.0. Rep. EPA/600/R-05/040, Environmental Protection Agency, 285 pp.

  • Salas, J. and Obeysekera, J. 2014: Revisiting the concepts of return period and risk for nonstationary hydrologic extreme events. J. Hydrol. Eng., 19, 554568, doi:10.1061/(ASCE)HE.1943-5584.0000820.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Schiermeier, Q., 2011: Increased flood risk linked to global warming. Nature, 470, 316316, doi:10.1038/470316a.

  • SCS, 1986: Urban hydrology for small watersheds. USDA TR-55, 164 pp. [Available online at https://www.nrcs.usda.gov/Internet/FSE_DOCUMENTS/stelprdb1044171.pdf.]

  • Seidou, O., A. Ramsay, and I. Nistor, 2012: Climate change impacts on extreme floods I: Combining imperfect deterministic simulations and non-stationary frequency analysis. Nat. Hazards, 61, 647659, doi:10.1007/s11069-011-0052-x.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Serinaldi, F., 2015: Dismissing return periods! Stochastic Environ. Res. Risk Assess., 29, 11791189, doi:10.1007/s00477-014-0916-1.

  • Smith, B. K., J. A. Smith, M. L. Baeck, and A. J. Miller, 2015: Exploring storage and runoff generation processes for urban flooding through a physically based watershed model. Water Resour. Res., 51, 15521569, doi:10.1002/2014WR016085.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Stedinger, J. R., and V. W. Griffis, 2011: Getting from here to where? Flood frequency analysis and climate. J. Amer. Water Resour. Assoc., 47, 506513, doi:10.1111/j.1752-1688.2011.00545.x.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Surfleet, C., and D. Tullos, 2013: Variability in effect of climate change on rain-on-snow peak flow events in a temperate climate. J. Hydrol., 479, 2434, doi:10.1016/j.jhydrol.2012.11.021.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Taylor, K. E., R. Stouffer, and G. Meehl, 2012: An overview of CMIP5 and the experiment design. Bull. Amer. Meteor. Soc., 93, 485498, doi:10.1175/BAMS-D-11-00094.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Tolson, B. A., and C. A. Shoemaker, 2007: Dynamically dimensioned search algorithm for computationally efficient watershed model calibration. Water Resour. Res., 43, W01413, doi:10.1029/2005WR004723.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Tramblay, Y., L. Neppel, J. Carreau, and K. Najib, 2013: Non-stationary frequency analysis of heavy rainfall events in southern France. Hydrol. Sci. J., 58, 280294, doi:10.1080/02626667.2012.754988.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Tramblay, Y., E. Amoussou, W. Dorigo, and G. Mahé, 2014: Flood risk under future climate in data sparse regions: Linking extreme value models and flood generating processes. J. Hydrol., 519, 549558, doi:10.1016/j.jhydrol.2014.07.052.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Trenberth, K. E., 2011: Changes in precipitation with climate change. Climate Res., 47, 123138, doi:10.3354/cr00953.

  • USGS, 2015: USGS 04234000 Fall Creek near Ithaca, NY. Accessed 11 November 2016. [Available online at http://waterdata.usgs.gov/nwis/uv?04234000.]

  • USGS, 2016: National elevation dataset. Accessed 11 November 2016. [Available online at http://nationalmap.gov/elevation.html.]

  • Walter, M., P. Gérard-Marchant, T. Steenhuis, and M. Walter, 2005: Closure to “Simple estimation of prevalence of Hortonian flow in New York City watersheds” by M. Todd Walter, Vishal K. Mehta, Alexis M. Marrone, Jan Boll, Pierre Gérard-Marchant, Tammo S. Steenhuis, and Michael F. Walter. J. Hydrol. Eng., 10, 168169, doi:10.1061/(ASCE)1084-0699(2005)10:2(169).

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Wehner, M. F., 2013: Very extreme seasonal precipitation in the NARCCAP ensemble: Model performance and projections. Climate Dyn., 40, 5980, doi:10.1007/s00382-012-1393-1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Weijs, S. V., N. van de Giesen, and M. B. Parlange, 2013a: Data compression to define information content of hydrological time series. Hydrol. Earth Syst. Sci., 17, 31713187, doi:10.5194/hess-17-3171-2013.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Weijs, S. V., N. van de Giesen, and M. B. Parlange, 2013b: HydroZIP: How hydrological knowledge can be used to improve compression of hydrological data. Entropy, 15, 12891310, doi:10.3390/e15041289.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Westra, S., I. Varley, P. Jordan, R. Nathan, A. Ladson, A. Sharma, and P. Hill, 2010: Addressing climatic non-stationarity in the assessment of flood risk. Aust. J. Water Resour., 14, 116.

    • Search Google Scholar
    • Export Citation
  • Wood, A. W., and J. C. Schaake, 2008: Correcting errors in streamflow forecast ensemble mean and spread. J. Hydrometeor., 9, 132148, doi:10.1175/2007JHM862.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Wood, A. W., T. Hopson, A. Newman, L. Brekke, J. Arnold, and M. Clark, 2016: Quantifying streamflow forecast skill elasticity to initial condition and climate prediction skill. J. Hydrometeor., 17, 651668, doi:10.1175/JHM-D-14-0213.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Wyżga, B., J. Zawiejska, and A. Radecki-Pawlik, 2016: Impact of channel incision on the hydraulics of flood flows: Examples from Polish Carpathian rivers. Geomorphology, 272, 1020, doi:10.1016/j.geomorph.2015.05.017.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Ye, H., E. J. Fetzer, A. Behrangi, S. Wong, B. H. Lambrigtsen, C. Y. Wang, J. Cohen, and B. Gamelin, 2016: Increasing daily precipitation intensity associated with warmer air temperatures over Northern Eurasia. J. Climate, 29, 623636, doi:10.1175/JCLI-D-14-00771.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Yossef, N. C., H. Winsemius, A. Weerts, R. Beek, and M. F. Bierkens, 2013: Skill of a global seasonal streamflow forecasting system, relative roles of initial conditions and meteorological forcing. Water Resour. Res., 49, 46874699, doi:10.1002/wrcr.20350.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • View in gallery

    Fall Creek watershed; the city of Ithaca; and locations of air temperature, snowpack depth, groundwater elevation, and streamflow monitoring stations.

  • View in gallery

    Model corroboration for Fall Creek (a) snowpack depth, (b) groundwater elevation, and (c) streamflow for 2013.

  • View in gallery

    Synthetic years of (a) lake elevation, (b) air temperature, (c) SWE, (d) groundwater elevation, (e) unsaturated zone soil moisture, (f) AET, and (g) the stream depth response to 1-yr/24-h precipitation event for each day.

  • View in gallery

    Two-dimensional distributions of flood stage for warm-season days; color scale indicates flood depth within Fall Creek (m).

  • View in gallery

    As in Fig. 4, but for cool-season days.

  • View in gallery

    Empirical probability density functions for daily changes to Cayuga Lake elevation, groundwater elevation, unsaturated zone soil moisture, snowpack, air temperature during warming period, and air temperature during cooling period.

  • View in gallery

    Flood hazard predictions as a function of hydrologic state during saturated conditions for past known conditions (blue) and future projected conditions (yellow). The dashed line indicates the upper bound of the 90% confidence interval.

  • View in gallery

    As in Fig. 7, but for drought conditions.

  • View in gallery

    Monthly distributions of hydrologic state variables and peak runoff for a fixed precipitation event under current [(a),(e),(i),(m) baseline] and projected [(b),(f),(j),(n) scenario A; (c),(g),(k),(o) scenario B; and (d),(h),(l),(p) scenario C] climate change conditions. Panels (n)–(p) show distributions of difference in flow from baseline case (m) (hypothetical scenario minus baseline).

  • View in gallery

    As in Fig. 9, but for scenarios D, E, and F.

All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 29 29 16
PDF Downloads 18 18 11

Hydrologic State Influence on Riverine Flood Discharge for a Small Temperate Watershed (Fall Creek, United States): Negative Feedbacks on the Effects of Climate Change

View More View Less
  • 1 Department of Biological and Environmental Engineering, Cornell University, Ithaca, New York
  • 2 Department of Earth and Atmospheric Sciences, Cornell University, Ithaca, New York
  • 3 Department of Biological and Environmental Engineering, Cornell University, Ithaca, New York
© Get Permissions
Full access

Abstract

Watershed flooding is a function of meteorological and hydrologic catchment conditions. Climate change is anticipated to affect air temperature and precipitation patterns such as altered total precipitation, increased intensity, and shorter event durations in the northeastern United States. While significant work has been done to estimate future meteorological conditions, much is currently unknown about future changes to distributions of hydrologic state variables. High-resolution hydrologic simulations of Fall Creek (Tompkins County, New York), a small temperate watershed (324 km2) with seasonal snowmelt, are performed to evaluate future climate change impacts on flood hydrology. The effects of hydrologic state and environmental variables on river flood stage are isolated and the importance of groundwater elevation, unsaturated soil moisture, snowpack, and air temperature is demonstrated. It is shown that the temporal persistence of these hydrologic state variables allows for an influence on watershed flood hydrology for up to 20 days. Finally, six hypothetical climate change forcing scenarios are simulated to estimate the influence of catchment conditions on the watershed runoff response. The possibility of drier summers and wetter springs with a reduced winter snowpack in the Northeast is also simulated. These hydrologic changes influence flood discharge in the opposite direction as climate effects because of a reduced snowpack accumulation and melt time. Strong hydrologic state influence on flood discharge may be most attributable to increased air temperature and decreased precipitation. Hydrologic state variables may change both the location and shape of seasonal flood discharge distributions despite expected consistency in the shape of precipitation statistic distributions.

© 2017 American Meteorological Society. For information regarding reuse of this content and general copyright information, consult the AMS Copyright Policy (www.ametsoc.org/PUBSReuseLicenses).

Corresponding author e-mail: James O. Knighton, james.knighton@gmail.com

Abstract

Watershed flooding is a function of meteorological and hydrologic catchment conditions. Climate change is anticipated to affect air temperature and precipitation patterns such as altered total precipitation, increased intensity, and shorter event durations in the northeastern United States. While significant work has been done to estimate future meteorological conditions, much is currently unknown about future changes to distributions of hydrologic state variables. High-resolution hydrologic simulations of Fall Creek (Tompkins County, New York), a small temperate watershed (324 km2) with seasonal snowmelt, are performed to evaluate future climate change impacts on flood hydrology. The effects of hydrologic state and environmental variables on river flood stage are isolated and the importance of groundwater elevation, unsaturated soil moisture, snowpack, and air temperature is demonstrated. It is shown that the temporal persistence of these hydrologic state variables allows for an influence on watershed flood hydrology for up to 20 days. Finally, six hypothetical climate change forcing scenarios are simulated to estimate the influence of catchment conditions on the watershed runoff response. The possibility of drier summers and wetter springs with a reduced winter snowpack in the Northeast is also simulated. These hydrologic changes influence flood discharge in the opposite direction as climate effects because of a reduced snowpack accumulation and melt time. Strong hydrologic state influence on flood discharge may be most attributable to increased air temperature and decreased precipitation. Hydrologic state variables may change both the location and shape of seasonal flood discharge distributions despite expected consistency in the shape of precipitation statistic distributions.

© 2017 American Meteorological Society. For information regarding reuse of this content and general copyright information, consult the AMS Copyright Policy (www.ametsoc.org/PUBSReuseLicenses).

Corresponding author e-mail: James O. Knighton, james.knighton@gmail.com

1. Introduction

Precipitation-induced watershed flooding is a function of the distribution of precipitation characteristics and the distribution of hydrologic catchment conditions, frequently referred to as hydrologic state variables. Increased global temperatures will raise the moisture holding capacity of the atmosphere and result in more intense precipitation events (Kunkel et al. 2013; Balling and Goodrich 2011; Wehner 2013; Hayhoe et al. 2007, 2008; Diffenbaugh et al. 2005). Recent interest in climate change–induced shifts in precipitation patterns has instilled a commonly held belief that flooding risk is increasing globally (e.g., Trenberth 2011; Schiermeier 2011). Trenberth (2011) and Schiermeier (2011) logically approach the problem of global riverine flood hazard from the point of view of precipitation forcing. Riverine flooding is commonly the result of the rainfall–runoff hydrologic response of the landscape, and it is therefore reasonable to assume that a shift toward more intense precipitation over a watershed will result in a uniform corresponding shift in the flood hazard regime toward increased flooding risk. Though we expect a net increase in global flooding, the increase in flooding may not be uniform. For example, Hirabayashi et al. (2013) demonstrate that there is a spatial heterogeneity of the anticipated change in flood risk, with some areas showing an expected decrease. Significant work has been done to estimate future meteorological conditions, though much is currently unknown about future changes to distributions of hydrologic state variables and how these changes in turn will influence riverine flood hydrology.

The runoff response of riverine systems is a constantly changing function of geomorphic (e.g., Wyżga et al. 2016), hydrologic (e.g., Nied et al. 2013), anthropogenic (e.g., Blöschl et al. 2015; Li et al. 2014), ecological (e.g., Butler 1989), and atmospheric (e.g., Wood et al. 2016) conditions. High-frequency (from daily to weekly) fluctuations in hydrologic states of a watershed have an influence on the runoff response to precipitation and therefore riverine flood discharge. Despite the potential for a rapidly varying hydrologic state to introduce additional variations in the rainfall–runoff watershed response, the importance of these conditions for making accurate discharge predictions remains the subject of some debate (Wood et al. 2016).

The concept of decadal nonstationary flood risk is receiving more attention as climate-induced changes in precipitation intensity and frequency threaten to increase flood hazard globally. Hirsch and Ryberg (2012) performed a statistical analysis to demonstrate that observed trends in riverine flow may not be linked to long-term changes in atmospheric CO2. Mallakpour and Villarini (2015) present an alternate conclusion where the frequency of riverine flooding events has increased under the driving assumption that an atmosphere at a higher temperature has resulted in increased precipitation. Bouwer (2011) and Blöschl et al. (2015) suggest that the observed global increase in flooding could be due to increased intense precipitation, though it is likely that land use and population changes have also modified the structure of risk. We propose that some of the disparity in these results is due to the differing hydrologic states of the watersheds considered. Hydrologic conditions are anticipated to respond to climate change and therefore have some inherent ability to transform the relationship between changing climate patterns and riverine flood hazard (e.g., Bell et al. 2016; Koplin et al. 2014; Blöschl et al. 2007).

Decadal shifts in flood hazard, such as those associated with climate change, are of concern for long-term urban planning (Gersonius et al. 2013; Kundzewicz et al. 2014; Jabareen 2013). Obeysekera and Salas (2014), Tramblay et al. (2014), Salas and Obeysekera (2014), Seidou et al. (2012), Westra et al. (2010), Serinaldi (2015), Gilroy and McCuen (2012), and Stedinger and Griffis (2011) present methodologies for assessing the potential for climate-induced nonstationary flood risk. Condon et al. (2015) and Tramblay et al. (2013) present case studies for future projected nonstationary flood hazard under climate change. These studies, while informative for planning, do not investigate the possibility that hydrologic state feedbacks on riverine flooding are occurring within the system, which may have the effect of further increasing or decreasing the distribution of flood hazard. Higher-frequency changes in flood discharge occur on seasonal as well as daily time scales. Hydrologic state variables (soil moisture, groundwater elevation, snowpack, and potential evapotranspiration) and environmental variables (temperature and downstream hydraulic boundary conditions) may have a significant effect on flood discharge over weekly to monthly time scales (Lo and Famiglietti 2010; Wood et al. 2016) and may themselves change in response to a changing climate (Blöschl et al. 2007).

The importance of hydrologic state and environmental variables for predicting rainfall–runoff responses has been studied extensively as a component of flood prediction systems. Wood et al. (2016) and Wood and Schaake (2008) suggest that these hydrologic state variables (i.e., initial conditions) may have a significantly lower effect on the skill of flood hazard predictions than the forecasted meteorological conditions. Alternately, Yossef et al. (2013) demonstrate that initial hydrologic conditions may be significant in certain types of watersheds. Hydrologic forecasts for basins influenced by snowmelt cycles were shown to be highly dependent on initial conditions of the snowpack. Yossef et al. (2013) also conclude that in larger basins the groundwater and unsaturated zone moisture content may have a noticeable effect on forecasted streamflow. Mahanama et al. (2012) demonstrate that knowledge of both snowmelt and soil moisture significantly improved model-based streamflow predictions in the U.S. Northeast.

Research objectives

First, we demonstrate which hydrologic state variables have an influence on the rainfall–runoff flood stage response of the Fall Creek watershed (Tompkins County, New York). Fall Creek is a relatively small basin (drainage area 324 km2) with a flood regime strongly influenced by spring precipitation and snowmelt events (USGS 2015). We propose that small watersheds with saturation-excess precipitation runoff responses and a significant snowfall and melt cycle, such as those found within Tompkins County, may be largely influenced by the present-day hydrologic state, similar to conclusions presented by Yossef et al. (2013) and Mahanama et al. (2012).

We further illustrate the sensitivity of riverine flood stage to hydrologic conditions by examining the temporal persistence of riverine flood stage with respect to extreme hydrologic initial conditions. Weijs et al. (2013a,b) demonstrate the tendency for hydrologic states (in their case, streamflow) to be highly compressible datasets due to a strong temporal persistence. They show that hydrologic time series can be described with less computer memory based on our knowledge of how present conditions are strongly related to past conditions. Hydrologic conditions often have a high autocorrelation, particularly during drought conditions. We propose that hydrologic state variables are not only a consideration for flood-initiating conditions at the onset of precipitation, but also introduce a form of memory to the rainfall–runoff response of a landscape.

Finally, we present estimates of the influence of anticipated climate change forcing (temperature and precipitation) on changes to the seasonal distributions of hydrologic state variables and, in turn, the flood regime. While our methodology is applied to a small temperate watershed with a seasonal snowmelt pattern, the approach is generalizable to any watershed type. We agree that more intense precipitation may result in a greater proportion of runoff, but will likely lower soil water and groundwater. Similarly, increased winter temperatures will decrease snowpack recharge. We hypothesize that the flood discharge from a small temperate watershed with a shallow confining layer and seasonal snow accumulation and melt may not necessarily increase with respect to climate change. We present research that considers only hydrologic state variables in isolation to determine the direction of the effect of this system component under hypothetical climate change forcing. We present detailed hydrologic modeling of a watershed forced with hypothetical climate change precipitation and temperature datasets to investigate hypothetical changes to the seasonal distributions of hydrologic state variables and subsequent effects on flood discharge.

2. Methodology

a. Study location: Fall Creek watershed

Fall Creek is a fourth-order stream that flows through Ithaca, New York, United States (Fig. 1). Within Ithaca, Fall Creek is contained within a man-made 30-m-wide trapezoidal earthen channel. Extreme precipitation events over the Fall Creek watershed have resulted in several flooding events within Ithaca in recent history (M. Thorn, city engineer, 2016, personal communication). Flooding of Fall Creek and the surrounding neighborhoods occurred in 2005 and 1993 (USGS 2015) as peak flows overwhelmed the capacity of the existing channel within Ithaca and overtopped the earthen levees. The flood elevation for Fall Creek is 3 m above the channel bed and stream gauge datum. All hazard estimates presented within this research are representative of the flow depth of Fall Creek at Ithaca.

Fig. 1.
Fig. 1.

Fall Creek watershed; the city of Ithaca; and locations of air temperature, snowpack depth, groundwater elevation, and streamflow monitoring stations.

Citation: Journal of Hydrometeorology 18, 2; 10.1175/JHM-D-16-0164.1

The contributing watershed to Fall Creek is a 324-km2 mixed forested and agricultural watershed located in Tompkins County (42°28′N, 76°27′W). The soil profile consists primarily of silt loam and silty clay loam, with a shallow confining layer at a depth of approximately 0.5–1 m (NRCS 2015). Easton et al. (2007) and Dahlke et al. (2009) propose that the rainfall–runoff response of New York primarily exists as a saturation-excess hydrologic condition. They propose that precipitation events typically do not generate overland runoff from most of the watershed and only do so when and where soils are highly saturated as a saturation-excess process.

Fall Creek drains to Cayuga Lake on the north side of Ithaca. The water surface elevation of Cayuga Lake varies seasonally and serves as a downstream boundary condition on Fall Creek. The elevation of Cayuga Lake is recorded at a daily time interval from 1956 to present (USGS 2015). Lake levels are the downstream boundary condition on Fall Creek. High lake levels could feasibly cause flooding during smaller precipitation events because of a higher hydraulic grade line (HGL) required to push flow out to Cayuga Lake against the boundary condition. Daily values are assumed constant over a given day. Hourly air temperature data for the Fall Creek watershed were obtained from the Ithaca Tompkins Regional Airport (NOAA/NCEI 2015). Hydrologic computations within the Storm Water Management Model (SWMM) interpolate between hourly air temperature values. Precipitation data for Fall Creek were obtained from a 60-yr record of hourly precipitation depths recorded at the Game Farm Road weather station in Tompkins County (42°27′N, 76°27′W) at an elevation of 292 m [North American Vertical Datum of 1988 (NAVD 88); NRCC 2015].

Fall Creek at Ithaca has a mean 15-min flow rate of 3.71 m3 s−1 (10th and 90th percentiles of 1.61 and 10.62 m3 s−1, respectively; USGS 2015). The watershed experiences mean annual precipitation of 94.7 cm (10th and 90th percentiles of 84.3 and 114.5 cm, respectively), with approximately 17% of the mean annual precipitation falling as snow (NRCC 2015). Air temperatures range from −11.6° to 17.2°C (10th and 90th percentiles of hourly air temperatures; NRCC 2015). Cayuga Lake water surface elevations range from 115.67 to 116.74 m [National Geodetic Vertical Datum of 1929 (NGVD 29); 10th and 90th percentiles of daily average lake elevation; USGS 2015].

b. Flood stage and discharge modeling with SWMM

We develop a U.S. Environmental Protection Agency (EPA) SWMM (build 5.1.010) of Fall Creek to predict hydrologic responses and flood stage and discharge. SWMM is a combined hydrologic (rainfall–runoff) and hydraulic (stream routing) model. The hydrologic module of SWMM is capable of predicting infiltration- and saturation-excess rainfall–runoff responses, unsaturated zone soil moisture fluctuations, groundwater flow, and snow accumulation and melt (Rossman 2010).

Snowmelt is calculated based on heat budget accounting via a modified degree-day method. Melt from the accumulated snowpack is determined from existing snowpack temperature and moisture content, energy inputs from air and precipitation, and user-supplied base temperature and melt coefficients scaled to the day of the year [for additional details on SWMM snowmelt, see Rossman (2010)]. We note that SWMM does account for mean watershed elevation in computing snowmelt, yet it does not account for differences in elevations between subcatchments that may have an influence on snowmelt. Grusson et al. (2015) demonstrate this effect of elevation on snowmelt within the Soil and Water Assessment Tool (SWAT).

Infiltration is solved with the Green and Ampt infiltration model. The groundwater module imposes saturation-excess conditions when the simulated water table rises to the ground surface. The unsaturated zone moisture content controls the instantaneous infiltration rate. SWMM is a lumped-term hydrologic model. Each subcatchment is defined by one set of representative parameter values (e.g., slope and soil textures). The hydraulic model of SWMM is one-dimensional routing model based on the dynamic-wave solution of the St. Venant equations.

The Fall Creek watershed model is composed of 424 individual subcatchments with a mean contributing area of 0.6 km2. The hydraulic model is a simplified drainage network composed of 363 conduits. Model topography and bathymetry were determined from the National Elevation Dataset (USGS 2016). We force the model with hourly precipitation and air temperature data and daily lake water surface elevations. The time step used for all hydrologic calculations (e.g., infiltration, runoff, and snowmelt) is 5 min. The hydraulic time step for in-channel routing was 30 s to meet the Courant–Friedrichs–Lewey condition of the explicit dynamic-wave solution technique.

c. Model corroboration

Brigode et al. (2014) suggest that the hydrologic model parameterization is a significant component of a probabilistic flood hazard prediction. They suggest that hydrologic calibration is potentially more important than initial hydrologic conditions in flood hazard estimation; however, they note that these findings may be erroneous in watersheds with a strong groundwater influence. We propose that the shallow confining layer of Tompkins County will result in hydrologic state variables having a significant influence on flood hazard. The influence of hydrologic parameterization on flood hazard estimates is acknowledged as important.

We calibrated our hydrologic model to observed daily snowpack, daily groundwater elevations, and hourly streamflow data from 1 January 2013 to 1 January 2014. By evaluating snowpack, groundwater, and streamflow, we may ensure that the Fall Creek SWMM adequately represents multiple components of the hydrologic cycle. Daily snowpack observations were obtained from the Ithaca Cornell University station (ID 304174; NOAA/NCEI 2015). Daily groundwater elevations were obtained from USGS groundwater station 422920076275301 (USGS 2016). We compare average watershed daily snowpack and groundwater depth to point measurements of each observation. We acknowledge that snowpack and groundwater depth observations were made outside of the Fall Creek watershed; however, we believe these are representative measurements because of the very close proximity (Fig. 1) and similarity in geology (NRCS 2015). Hourly streamflow data for the Fall Creek watershed were obtained from USGS gauge 04234000 (USGS 2015).

We perform parameter calibration using the dynamically dimensioned search (DDS) hydrologic parameter calibration algorithm (Tolson and Shoemaker 2007) with the Nash–Sutcliffe efficiency (NSE; Nash and Sutcliffe 1970) as the model performance objective function. In the case of the snowpack model we use two objective functions, NSE and percentage of days in which the model correctly simulates snow presence versus absence (P/A) to avoid allowing large pack accumulations from dominating the NSE and therefore the snowpack calibration. For each submodel (snow, groundwater, and streamflow), we ran the DDS algorithm for 1000 simulations with a disturbance factor of 0.2 to identify the region of the optimal parameter set. We then perform a second set of 1000 simulations with a disturbance factor of 0.05 to refine the model parameters. SWMM parameters’ feasible ranges supplied to the DDS algorithm values are presented in Table 1.

Table 1.

SWMM calibration parameter ranges and final calibrated values.

Table 1.

d. Flood stage response to hydrologic state variables

1) Development of initial hydrologic conditions

We first estimate the riverine flood stage response with respect to naturally varying hydrologic conditions. We utilize artificial combinations of annual forcing time series as opposed to chronologically observed data to simulate more extreme combinations of temperature, precipitation, and lake elevation than have been observed throughout the existing period of record.

We generate synthetic 10-yr time series of daily lake elevation and hourly air temperature data by resampling from the historical observed records of each dataset. This sampling procedure was chosen to preserve the temporal autocorrelation among variables. Variables are resampled independent of each other. Corresponding daily actual evapotranspiration (AET) values were simulated from daily minimum and maximum temperatures (Archibald and Walter 2014; Fuka et al. 2013).

We develop a corresponding 10-yr synthetic record of hourly precipitation representing present-day meteorological conditions through copula modeling of the dependence of precipitation event statistics. Our methodology is similar to those presented in Haberlandt and Radtke (2014) and Paschalis et al. (2014). Genest and Favre (2007) present a review of the application of the copula concept within hydrology. We first develop probability distributions of the precipitation event statistics of depth, peak intensity, duration, temporal loading, and interevent time from the observed 60-yr precipitation record. We then fit a t-copula model to the conditional event statistics of depth, intensity, duration, and temporal loading. The copula model and marginal distributions for duration and interevent time are then randomly sampled to generate synthetic precipitation event statistics. Further details of precipitation generation methodology are outlined in Knighton and Walter (2016). We use these statistics to construct a synthetic hourly time series of precipitation.

We apply these synthetic-forcing-data time series to the SWMM to develop daily estimates of subcatchment groundwater elevation, unsaturated zone soil moisture content, and snow water equivalent (SWE) for each subcatchment to be used as initial hydrologic conditions.

2) Sensitivity of flood stage response to hydrologic state variables

Next, we perform a second series of simulations to determine what the runoff response to a fixed amount of precipitation would have been on each day over the 10-yr synthetic period. In this way, we may determine how changing hydrologic state variables affects the flood stage response independent of the probability of precipitation. Our methodology is similar to the “long-term hydrologic simulation method” for flood hazard responses described in Lawrence et al. (2014). Lawrence et al. (2014) use this methodology to estimate the distribution of aleatory uncertainty surrounding the rainfall–runoff response for a given design precipitation event. While our methodology is similar, we employ this procedure to evaluate both uncertainty surrounding flood depth and runoff responses to the 1-yr precipitation event as well to study relationships between hydrologic state variables and flooding potential.

Hydrologic state variables for each day are initialized using the daily hydrologic conditions determined in section 2c(1). Next, we replace the actual precipitation with a synthetic 44.6-mm precipitation event occurring over 24 h, distributed based on the Soil Conservation Service (SCS) Type-II hyetograph (SCS 1986). This event is approximately the 1-yr storm for Ithaca as estimated by NOAA/HDSC (2016). The somewhat arbitrary precipitation forcing event was chosen to be large enough that infiltration-excess runoff would occur during periods of drought and small enough that saturation-excess conditions would not dominate each simulation independent of the initial hydrologic conditions. Each fixed precipitation simulation was carried out for 3 days to capture all flood routing within Fall Creek. Following each simulation, initial hydrologic conditions are reset to those determined in section 2c(1) and the design precipitation is advanced forward one day. This procedure is repeated throughout the entire time series.

We divide the dataset into “warm” and “cool.” The dividing dates between the datasets are 1 May and 1 October. First- and second-order interactions between hydrologic state variables and flood stage are evaluated. We perform a regression analysis of each initial condition hydrologic state variable against the resulting peak water surface elevation of Fall Creek within Ithaca as a measure of the flood hazard. We use the linear regression coefficient of determination R2 and Spearman’s ranked correlation coefficient ρ as measures of the importance of this particular state variable on flood stage prediction.

e. Flood stage persistence due to hydrologic states

We evaluate how much future hydrologic flood regime is controlled by present-day hydrologic conditions. Our hypothesis is that hydrologic extremes will influence the rainfall–runoff response for a significant period of time. Previous research has demonstrated that hydrologic state variables may be used to refine predictions of future discharge (Wood et al. 2016; Yossef et al. 2013; Mahanama et al. 2012; Lo and Famiglietti 2010). Similarly, Weijs et al. (2013a,b) demonstrate that this prediction capability exists because of the temporal persistence of hydrologic states. We intend to demonstrate this temporal influence by performing random walk forecasts beginning from two hydrologic extremes: “saturated” and “drought.”

To demonstrate this effect, we select 2 days within the synthetic 10-yr continuous record to represent present-day conditions under high and low runoff potential. We select the high runoff day (herein referred to as the “saturation” case) as a day immediately following a large precipitation recharge event with a significant snowpack during the spring season where the ambient air temperature regularly crosses the dividing temperature between snow and rain. We select the low runoff day (herein referred to as the “drought” case) as a day with seasonally low water table and unsaturated zone moisture content during the summer months with no snowpack. We present these two case studies to demonstrate not only the immediate effect of hydrologic state variables on flood risk, but the ability for extreme hydrologic conditions of the landscape to control the persistence of flood stage.

For each case, we project future hydrologic and environmental conditions by allowing hydrologic (snowpack, soil moisture, and groundwater) and environmental (lake elevation) variables to progress along a random walk [Eq. (1)]. We define the random walk step length as the change in hydrologic state value from the first time step of each day. The distribution defining each random walk step length was derived from the time series of each hydrologic state variable as determined in section 2c(1):
e1
where Ht is the hydrologic or environmental state at time t and φt is the random disturbance drawn from distribution of hydrologic state variable daily step lengths.
Temperatures were modeled as a random walk with mean reversion to the monthly mean temperature:
e2
where Tt is air temperature at time t, φt is random disturbance drawn from distribution of air temperature daily step lengths, β is the autoregression coefficient, and μ is distribution mean (monthly average temperature).

The autoregression coefficient β derived from the historical daily temperatures was 0.9116. Temperature random walks were permitted to vary between −31° and 27°C based on historically observed extremes. Temperatures that fell outside of these thresholds were resampled. Random walks days with above-freezing temperatures have the snowpack depth reduced based on the modified degree-day method of SWMM (described in section 2b).

For each random walk for each hydrologic state, we simulate the 1-yr precipitation event at each day to determine the distribution of flood stage responses of Fall Creek. For each case, we compute 100 parallel walks each with a random walk length of 100 days.

f. Future flood frequency estimation

1) Development of subdaily future meteorological forcing data

We develop hypothetical future climate forcing data for the Fall Creek watershed. We review seven CMIP5 multimodel ensemble projections for the year 2100 (Taylor et al. 2012; Table 2) for Representative Concentration Pathway 8.5 (RCP 8.5) to bracket the potential ranges considered for probable future annual precipitation and daily minimum and maximum temperatures.

Table 2.

CMIP5 multiensemble global climate models used to estimate future changes to annual precipitation and daily max and min temperatures. The ΔP column provides the estimated difference in median annual precipitation between the periods of 2015–25 and 2080–99. The ΔT column provides the estimated difference in median annual max air temperature between the periods of 2015–25 and 2080–99.

Table 2.

CMIP5 RCP 8.5 multimodel ensemble projections show agreement in the median annual maximum and minimum daily temperature increase of approximately 3°–5°C by 2100 (Table 2). We develop sets of synthetic hourly air temperature forcing data based on the ad hoc assumption that future annual temperatures can be represented by historical temperature time series shifted up by 2°, 3°, and 4°C. Though more significant changes in the ambient air temperature in the form of extreme temperature events are likely to occur (Diffenbaugh et al. 2005), we adopt this simplifying assumption as we are primarily interested in examining the effects of increased time spent above 0°C.

CMIP5 multimodel ensemble projections demonstrate a range of future annual average precipitation for Ithaca from no substantial change (ACCESS1.3, BCC_CSM1.1, BNU-ESM, CanESM2, CCSM4, CNRM-CM5, and GFDL-ESM2G) to a decrease (INM-CM4.0) in median annual precipitation totals by 2100 (Table 2). CMIP5 projections include daily total precipitation that does not properly inform us on flooding responses within smaller catchments that are possibly controlled by subdaily duration precipitation events (Knighton and Walter 2016). To develop subdaily precipitation time series, we manipulate the copula rainfall distribution [described in section 2c(1)] parameters of peak intensity and event depth. DeGaetano (2009) provides evidence that the distributions of extreme precipitation events are shifting toward more intense events. DeGaetano (2009) shows that the generalized extreme value (GEV) distribution’s describing precipitation event statistics display a change in the distribution location parameter, but consistency in the shape and scale parameter. We propose therefore that the shape of distributions of rainfall event statistics determined for Tompkins County can be maintained and the location parameters may be modified to simulate changes to future precipitation patterns based on climate change.

Palecki et al. (2005) present changes in the northeastern U.S. 15-min precipitation patterns through a cluster analysis of records from 1972 to 2002. They present observed changes to distributions of the precipitation event statistics depth, average intensity, peak intensity, and event duration. These published values suggest that peak 15-min intensity is generally increasing, whereas total depth and duration of precipitation are decreasing. Alternately, Diffenbaugh et al. (2005) present global climate simulations that suggest total precipitation depth is likely to remain constant under climate change; however, we will still see some increase in event intensity. These results are generally in agreement with the range of future annual precipitation presented in the CMIP5 multimodel ensemble projections. As discussed in Diffenbaugh et al. (2005), the interevent frequency of precipitation events in the Northeast is not likely to change under the effects of a changing climate. We therefore propose no modification to the interevent parameter of the rainfall event statistics.

We develop six ad hoc climatic forcing scenarios (Table 3) to capture a wide range of possible shifts in hydrologic forcing. We propose scenarios A, B, and C, which assume that total precipitation decreases while event average and peak intensity increase following observed trends identified by Palecki et al. (2005), Frumhoff et al. (2007), and Hayhoe et al. (2008, 2007) and CMIP5 model INM-CM4.0. Scenarios D, E, and F consider average and peak precipitation intensity to increase, while total precipitation depth remains constant as in Diffenbaugh et al. (2005) and CMIP5 models ACCESS1.3, BCC_CSM1.1, BNU-ESM, CanESM2, CCSM4, CNRM-CM5, and GFDL-ESM2G. Similarly, Ye et al. (2016) present findings for Eurasia that suggest higher atmospheric temperatures may be associated with an increasing in event intensity, but not a change to event depth.

Table 3.

Hypothetical climate change scenarios used to force the hydrologic model to determine changes in hydrologic state variables and flood hazard. The change in instantaneous ambient air temperature from historical observed temperatures is given in the second column. The peak hourly average precipitation intensity is given in the last column.

Table 3.

2) Landscape flood discharge response to changes in climatic forcing

Next, we isolate the effects of watershed hydrology on the future flood discharge estimates. We evaluate the effects of changes to temperature and precipitation forcing data on high-frequency changes in hydrologic state variables and in turn the effect this has on riverine flood discharge. We use the hypothetical future climate data [section 2d(1)] to force the Fall Creek SWMM to develop time series of hydrologic state variables under climate change. These hydrologic time series (soil moisture, groundwater elevation, and SWE) are then used to initialize simulations as described in section 2c(2) to determine the flood discharge response to a fixed precipitation event.

We evaluate the changes in flood discharge response to the present-day 1-yr/24-h precipitation event attributable to hydrologic state variables as outlined in sections 2d(1) and 2d(2). This analysis methodology specifically does not present the anticipated changes in flood discharge probability distributions, but rather isolates the effects of altered hydrologic state variables on flood discharge. Therefore, we do not modify the depth of 1-yr precipitation event between cases evaluated. By holding the design precipitation constant, we assess how the landscape response to climate change could affect flood discharge, focusing specifically the direction of the influence.

3. Results and discussion

a. Model corroboration

We corroborated the Fall Creek SWMM by simulating from 1 January 2013 to 1 January 2014 with observed hourly precipitation and temperature data and daily lake water surface elevations. The snowmelt model (NSE = 0.41, P/A = 89%), groundwater (NSE = 0.47), and streamflow (hourly NSE = 0.45, daily NSE = 0.5; Table 1, Fig. 2) are adequate representations of the hydrology (Moriasi et al. 2007). For the snowpack model, we note that some observed trace snowpack accumulations were not simulated; however, the major snowmelt events of 2013 are captured. SWMM adequately represents both winter and summer groundwater and streamflow dynamics, specifically, event peak flow rates (Fig. 2). We note, however, as stated in Blöschl et al. (2007), that our model is imperfect and that this imperfect representation of hydrology likely has some effect on our results.

Fig. 2.
Fig. 2.

Model corroboration for Fall Creek (a) snowpack depth, (b) groundwater elevation, and (c) streamflow for 2013.

Citation: Journal of Hydrometeorology 18, 2; 10.1175/JHM-D-16-0164.1

b. Flood stage response to hydrologic state variables

The continuous synthetic record of hydrologic state variables was generated and simulated with the Fall Creek EPA SWMM. Figure 3 presents the 10 simulated years of hydrologic states and flood stage response to the 1-yr storm. The rainfall–runoff flood stage response to the 1-yr precipitation event over a 2-yr period varies considerably from 0.1 to 4 m (Fig. 3). Simulated runoff responses during summer months show a consistent flood stage of approximately 1–1.5 m within Fall Creek. During winter months the flood stage was less predictable, with alternating days of low and high stage. This result suggests a more complicated runoff response to hydrologic conditions during the cool season.

Fig. 3.
Fig. 3.

Synthetic years of (a) lake elevation, (b) air temperature, (c) SWE, (d) groundwater elevation, (e) unsaturated zone soil moisture, (f) AET, and (g) the stream depth response to 1-yr/24-h precipitation event for each day.

Citation: Journal of Hydrometeorology 18, 2; 10.1175/JHM-D-16-0164.1

The periods of the year showing the highest influence of hydrologic state variables on flood stage are the spring and fall seasons. During these periods of time, an elevated water table, soil moisture, potential presence of a snowpack, and temperatures crossing the precipitation/snow dividing boundary create a potential for large combined runoff and snowmelt events.

The effect of each hydrologic state variable on the watershed flood response was estimated using first-order interactions (Table 4) and then by considering two-dimensional hydrologic state regressions against the magnitude of the flood stage (Figs. 4, 5). We consider these two-dimensional representations of the results as they allow evaluation of the effects of marginal distributions as well as interactions between variables.

Table 4.

Correlation of hydrologic state variables and flood hazard as measured by coefficient of determination for a linear relationship and Spearman’s ranked correlation coefficient.

Table 4.
Fig. 4.
Fig. 4.

Two-dimensional distributions of flood stage for warm-season days; color scale indicates flood depth within Fall Creek (m).

Citation: Journal of Hydrometeorology 18, 2; 10.1175/JHM-D-16-0164.1

Fig. 5.
Fig. 5.

As in Fig. 4, but for cool-season days.

Citation: Journal of Hydrometeorology 18, 2; 10.1175/JHM-D-16-0164.1

During warm-season days, the flood stage is primarily a function of the groundwater elevation and soil moisture content [Table 4, Fig. 4 (top left)]. The Fall Creek watershed does experience a saturation-excess response (Walter et al. 2005; Easton et al. 2007); however, it should be noted that the results suggest that infiltration-excess conditions throughout the watershed are also a consideration for predicting flood stage responses from larger precipitation events, particularly during the summer months. The highest flooding events occur during periods of high groundwater elevation and high soil moisture content.

Within natural watersheds, the soil moisture content at the start of precipitation has been shown to be a significant consideration for estimating the runoff response of a watershed (Massari et al. 2014a,b, 2015; Nied et al. 2013; Mahanama et al. 2012; Norbiato et al. 2008; Ravazzani et al. 2007; Smith et al. 2015; Elbialy et al. 2014). The Simulation Climato-Hydrologique pour l’Appréciation des Débits Extrêmes (SCHADEX) program is an example of a flood prediction system that provides estimates of future flood hazard based in part on historical hydrologic states including soil moisture content (Paquet et al. 2013). Soil profiles with a shallow confining layer exhibit a precipitation runoff response that is dependent on the depth to water table (Lo and Famiglietti 2010). Under infiltration-excess conditions, soil moisture controls the infiltration capacity and storage within the unsaturated zone. Within Tompkins County, a large proportion of runoff is related to saturation-excess conditions, which persist when the water table rises close to the land surface (Easton et al. 2007; Dahlke et al. 2009). An elevated water table reduces the total storage volume in the soil unsaturated zone, which would otherwise buffer additional rainfall or snowmelt depths. When the water table intersects the ground elevation (or nearly so), an impermeable surface is created that converts all rain to surface runoff. We therefore expect that the elevation of the water table should be a significant consideration for flood hazard in regions with shallow confining layers. Our results are therefore consistent with previous research.

Cayuga Lake serves as the downstream hydraulic boundary condition on Fall Creek and therefore may have some effect on the water surface elevation within the regionally relatively heavily populated areas of Ithaca (Fig. 4). The results of the flood stage regression suggest that the lake water surface elevation has a minimal impact on the peak flood elevation within Fall Creek, as the channelized portion within Ithaca is the hydraulically limiting element. The highest-magnitude floods do occur at the high lake elevations; however, we also see high-magnitude floods at lower lake elevations [Fig. 4 (top right)]. We therefore consider lake elevation to be a relatively nonsensitive state variable.

The instantaneous evapotranspiration of the landscape had no discernible influence on the precipitation–runoff response or the flood stage [Table 4, Fig. 4 (bottom right)]. The precipitation rate was significantly larger than the rate of AET and overwhelmed the effects of varying potential evapotranspiration (PET) losses during runoff. We note, however, that as AET has an influence on the unsaturated zone soil moisture, it is likely important over longer time scales.

Watersheds that experience freezing temperatures are subject to precipitation falling as snow and the accumulation of snowpacks. Days with warmer temperatures and increased solar radiation can induce greater runoff through snowmelt events. The extent and depth of the accumulated snowpack coupled with energy inputs has an influence on streamflow (Harrison and Bales 2015; Perju et al. 2013; Ceppi et al. 2013; Freudiger et al. 2014; Lawrence et al. 2014; Mahanama et al. 2012; Jörg-Hess et al. 2015). Ceppi et al. (2013) present a compelling case for evaluating the uncertainty in ambient air temperatures as well as precipitation when forecasting flood hazard to identify the proportion of precipitation partitioned as snow. Lawrence et al. (2014) expand on the work of Paquet et al. (2013) to introduce the effects of snow accumulation and melt coupled with extreme precipitation to improve prediction skill. During the cool season, flood stage is primarily correlated with air temperature [Fig. 5 (bottom left)]. The days producing the greatest runoff are days near 0°C. Air temperatures near 0°C ensure an accumulated snowpack close to melting. High discharge occurs across a wide range of groundwater and snowpack depths. This result suggests that while peak river stage is much more variable during the cool season, it remains a function primarily of air temperature.

During periods with an accumulated snowpack, the influence of hydrologic state variables becomes less clearly defined overall (Table 4, Fig. 5). The groundwater elevation and soil moisture have some influence on the flood stage. The ambient temperature has some influence as shown by the reduced flood stage responses for temperatures below 0°C. There is no discernible influence of the size of the accumulated snowpack on the flood stage (Table 4, Fig. 5). This result suggests that while snowpack melt does occur and has some influence on flood stage, the proportion of the snowpack that is melted on a given day is likely less than the total accumulated snowpack. We conclude that the only consideration with respect to snowpack is presence versus absence.

c. Flood stage persistence due to hydrologic states

Next, we demonstrate the temporal influence of extreme hydrologic state conditions with respect to flood stage. We develop probability distributions of changes in hydrologic state and environmental variables (Fig. 6) based on the 10-yr continuous simulation described in section 2c(1). These distributions demonstrate that changes to most hydrologic variables over a 24-h period are fairly constrained. The lake water surface elevation changes gradually because of the large buffering capacity of the lake relative to the volume of runoff and precipitation inputs. Groundwater elevation and soil moisture content similarly have a high temporal persistence. The temporal persistence of these hydrologic state variables provides us with some ability to then predict, with some confidence, the probability of future hydrologic states, and therefore future runoff potential.

Fig. 6.
Fig. 6.

Empirical probability density functions for daily changes to Cayuga Lake elevation, groundwater elevation, unsaturated zone soil moisture, snowpack, air temperature during warming period, and air temperature during cooling period.

Citation: Journal of Hydrometeorology 18, 2; 10.1175/JHM-D-16-0164.1

Temperature and snowpack demonstrate less temporal persistence. Large changes in the mean daily temperature result in significant melt events. Similarly, temperatures dropping below 0°C result in snowpack accumulation events as all precipitation contributes to the snowpack. We considered the distribution of daily mean temperature changes based on separate warming (from 1 February to 1 August) and cooling seasons (from 1 August to 1 February).

The random walks of each hydrologic state were considered to be independent, with the exception of the effect of temperature on snowpack accumulation as discussed in section 3b. We perform hypothetical flood stage projections for a “saturated” and “unsaturated” case (section 2d).

Figure 7, representing the saturated case, illustrates how the upper bound of the 90% confidence interval for flood stage remains elevated for a period of approximately 20 days following a large snowmelt and groundwater recharge event. Beyond 20 days, the flood stage distribution returns to the long-term average. The high variability immediately following the event is the result of temperatures above zero jointly occurring with a snowpack. Random walks that develop positive temperatures eventually deplete the snowpack. Random walks with below 0°C temperature build up the snowpack, but contribute less runoff. As the random walks diverge from the dividing temperature for precipitation type, the probability of such an extreme runoff–snowmelt event decreases. The upper envelop of flood stage estimates shows potentially large floods of up to 5 m that significantly decrease in likelihood beyond 20 days. This result further demonstrates temperatures near 0°C as being correlated with large flood potential, as in section 3b.

Fig. 7.
Fig. 7.

Flood hazard predictions as a function of hydrologic state during saturated conditions for past known conditions (blue) and future projected conditions (yellow). The dashed line indicates the upper bound of the 90% confidence interval.

Citation: Journal of Hydrometeorology 18, 2; 10.1175/JHM-D-16-0164.1

The “drought conditions” scenario (Fig. 8) shows a similar temporal persistence of the flood stage. The absence of an initial snowpack simplifies the predictions and reduces the large variance in flood stage seen in the “saturated condition” simulations.

Fig. 8.
Fig. 8.

As in Fig. 7, but for drought conditions.

Citation: Journal of Hydrometeorology 18, 2; 10.1175/JHM-D-16-0164.1

Understanding the present-day hydrologic state of a saturation-excess landscape with a snow season, like Fall Creek, provides considerable refinement of expected flood stage for an extended period of time (Figs. 7, 8). We demonstrate that hydrologic state variables act as a source of memory for systematic predictions of flood stage responses. With respect to a monthly flood stage prediction, these higher-frequency changes in the landscape result in a somewhat variable flood stage response. Hydrologic initial conditions beyond 20 days are generally unpredictable and therefore result in large uncertainty in a flood stage forecast. When viewed at a daily to weekly time scale, however, these gradually changing hydrologic conditions can have a lasting influence on the rainfall–runoff response of a watershed. We demonstrate how extreme hydrologic conditions affect not only the runoff response of the next precipitation event [as demonstrated in Nied et al. (2013)], but may constrain the possible range of the rainfall–runoff response for several weeks throughout multiple precipitation events.

d. Landscape flood discharge response to changes in climatic forcing

The hypothetical climate datasets [described in section 2c(1)] were used to force the Fall Creek watershed model to estimate responses in hydrologic state variables. We then introduce a design precipitation event on each day initialized by the long-term hydrologic state [as described in sections 2c(1), 2c(2), and 2e] to determine what effect climate change may have on the hydrologic state variables and how in turn this affects the aleatory uncertainty of the 1-yr precipitation event runoff response of Fall Creek.

Scenarios A, B, and C assume air temperatures increase, total precipitation depth decreases, and average and peak intensity increase as in Palecki et al. (2005), Frumhoff et al. (2007), and Hayhoe et al. (2008, 2007). Increasing temperature significantly decreases the accumulated snowpack during winter months (Figs. 9a–d). The anticipated reduced snowpack accumulation over the winter results in less spring melt runoff and groundwater recharge. Groundwater elevation and soil moisture similarly trend toward drier average conditions, with the exception of March, where earlier snowmelt due to higher temperatures results in increased ground saturation. We observe lower or unchanged median peak discharge for all months except March (Figs. 9n–p, Table 5). The hydrology of the peak flood controlling month, April, is strongly influenced by spring precipitation and snowmelt events. The reduced April snowmelt results in a general decrease in annual peak floods. Similarly, in summer the effects of climate change on hydrologic state variables indicate that an overall negative feedback on flood discharge is likely to be created similar to conclusions presented by Surfleet and Tullos (2013). The watershed soils become drier for greater periods of the year due to higher winter temperatures, lower total precipitation, and increased average and peak intensities and therefore have a greater infiltration potential.

Fig. 9.
Fig. 9.

Monthly distributions of hydrologic state variables and peak runoff for a fixed precipitation event under current [(a),(e),(i),(m) baseline] and projected [(b),(f),(j),(n) scenario A; (c),(g),(k),(o) scenario B; and (d),(h),(l),(p) scenario C] climate change conditions. Panels (n)–(p) show distributions of difference in flow from baseline case (m) (hypothetical scenario minus baseline).

Citation: Journal of Hydrometeorology 18, 2; 10.1175/JHM-D-16-0164.1

Table 5.

Median difference in 1-yr precipitation peak discharge (m3 s−1) between hypothetical climate change scenarios and baseline conditions. The values represent the hypothetical scenario minus the baseline, where a negative indicates a decrease in future flow that is attributable to the hydrologic state changes.

Table 5.

Scenarios D, E, and F assume that air temperatures and average and peak intensity increase, while total depth remains constant, as in Diffenbaugh et al. (2005) and Ye et al. (2016). As in scenarios AC, we see a reduced snowpack (Figs. 10a–d), wetter March, and drier April–September (Figs. 10e–l); however, the overall change in each hydrologic state variable from the baseline condition is significantly less (Figs. 9, 10). The runoff response of the watershed for scenarios D, E, and F shows a similar decrease (Figs. 10m–p, Table 5); however, the peak flow month, April, does not decrease as significantly as in scenarios AC. This result suggests that the air temperature and precipitation characteristic of total rainfall event depth may be more influential than peak hourly intensity in defining future hydrologic changes. We observe that the seasonality of the groundwater elevation is somewhat exaggerated in scenarios DF. Summer months become drier; however, spring and fall remain similar to baseline conditions.

Fig. 10.
Fig. 10.

As in Fig. 9, but for scenarios D, E, and F.

Citation: Journal of Hydrometeorology 18, 2; 10.1175/JHM-D-16-0164.1

We demonstrate in section 3b that initial hydrologic state conditions have a significant effect on flood response of the Fall Creek watershed to a design precipitation event. We then demonstrate in section 3c that hydrologic conditions introduce a form of memory in flood stage estimation that may persist up to 20 days beyond extreme conditions. This memory further explains why climate change–induced alterations of the hydrologic cycle within the Northeast may have an important negative feedback effect on flood hazard. Periods of reduced groundwater recharge (i.e., drought) influence not only the runoff response for the next precipitation event, but last until a significant precipitation depth has fallen to replenish the depleted reservoirs within the soil layers.

We present evidence that climate change may not necessarily increase flood discharge within Fall Creek despite an expected increase in intense precipitation, which is similar to regional conclusions presented by Hirabayashi et al. (2013). Changes to the hydrologic states show the response of the landscape is potentially trending toward a decreased rainfall–runoff response under climate change meteorological forcing. We agree that increased air temperatures and increased hourly precipitation intensity should, in general, result in increased flooding, as is proposed globally in Trenberth (2011) and Schiermeier (2011); however, we present evidence that allows us to refine this conclusion for small snowmelt influence watersheds with shallow confining layers.

Blöschl et al. (2007) hypothesize that feedbacks on hydrologic systems will form at the catchment scale as a result of modifying climate forcing. We observe this concept for Fall Creek in the form of reduced snowpack accumulation and soil saturation. Probability distributions for groundwater elevation and soil moisture content respond differently under the climate change forcing (Fig. 9). While both variables trend toward drier conditions, we see a reduced variance in groundwater elevation and an increased variance in soil moisture content across all months for scenarios AC (Fig. 9). Similarly, the shape and location of the probability distributions of watershed runoff from the 1-yr rainfall changes for each month through all cases evaluated. We observe not only reduced median peak runoff, but a reduced variance in the monthly distributions of the design precipitation runoff response. Our results suggest that feedbacks from hydrologic state variables may change both the location and shape of flood discharge probability distributions, despite consistency in the shape of precipitation event statistic distributions (DeGaetano 2009).

4. Conclusions

  1. We demonstrate that the hydrologic state variables of temperature, SWE, unsaturated zone soil moisture content, and groundwater elevation have an effect on flood stage using a long-term flood hazard methodology similar to that described in Lawrence et al. (2014). Flood stage in systems like Fall Creek is largely dominated by unsaturated zone soil moisture and groundwater elevation. Despite a more complex flooding response during the cool season months, we do not observe strong second-order interactions among hydrologic state variables and flood stage.
  2. We demonstrate the effects of extreme hydrologic state conditions on the temporal persistence of flood stage. From a fixed point in time, we estimate future probable hydrologic states through random walk modeling. Each random walk chain is simulated to estimate the probability distribution of flood stage runoff response for a fixed precipitation depth. Flood stage predictions for “saturated” and “drought” conditions evaluated each show a similar flood stage persistence of about 20 days before flood stage returns to a long-term average. The strong temporal persistence of hydrologic state variables causes a medium-range (20 day) constraint of possible flood stage. These findings further demonstrate the importance of hydrologic state variables for defining the rainfall–runoff response of a watershed.
  3. Watershed flooding is a function of both precipitation and hydrologic catchment conditions. Climate change is anticipated to result in increased air temperatures and altered precipitation patterns. While significant work has been done to estimate future meteorological conditions, much is currently unknown about future changes to distributions of hydrologic state variables and how they will affect the future flood regime. We demonstrate that watershed hydrology can limit or decrease flood discharge through hypothetical climate change scenarios. Higher summer air temperatures, higher precipitation intensity, and less total precipitation results in higher evaporative demand and less soil and groundwater recharge. Similarly, higher winter air temperatures result in a reduced snowpack accumulation and earlier melt. Decreased soil saturation and snowpack accumulation results in a lower runoff potential and therefore imposes a negative influence on flood discharge. While changes in precipitation patterns may be pushing toward increasing flood hazard, watershed hydrology likely has some ability to buffer this change.
  4. Hypothetical climate change scenarios in which we assume that total depth decreases and peak hourly intensity increases show a significantly larger reduction in the flood discharge than scenarios in which we assume only peak hourly intensity increases. These results suggest negative feedbacks on flood discharge may be more attributable to changes in air temperature and total precipitation event depth as opposed to only increased precipitation intensity.
  5. Blöschl et al. (2007) hypothesize that feedbacks on hydrologic systems will form at the catchment scale as a result of modifying climate forcing, which we observe for Fall Creek in the form of reduced snowpack accumulation and soil saturation. Our hypothetical climate change scenarios demonstrate that hydrologic state responses to climate change occur in both the location and shape of monthly distributions of soil moisture content and groundwater elevation. These results suggest that feedbacks from hydrologic state variables may change both the location and shape of flood discharge probability distributions, despite evidence for consistency in the shape of precipitation event statistic distributions (DeGaetano 2009).

Acknowledgments

We thank Michael Thorne (City of Ithaca, New York) and William Coon (U.S. Geological Survey) for their feedback on this work. We acknowledge the Northeast Regional Climate Center (NRCC) for providing precipitation observations. We thank three anonymous reviewers for their thoughtful comments.

REFERENCES

  • Archibald, J. A., and M. T. Walter, 2014: Do energy‐based PET models require more input data than temperature‐based models?—An evaluation at four humid FluxNet sites. J. Amer. Water Resour. Assoc., 50, 497508, doi:10.1111/jawr.12137.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Balling, R. C., Jr., and G. B. Goodrich, 2011: Spatial analysis of variations in precipitation intensity in the USA. Theor. Appl. Climatol., 104, 415421, doi:10.1007/s00704-010-0353-0.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Bell, V., A. Kay, H. Davies, and R. Jones, 2016: An assessment of the possible impacts of climate change on snow and peak river flows across Britain. Climatic Change, 136, 539553, doi:10.1007/s10584-016-1637-x.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Blöschl, G., and Coauthors, 2007: At what scales do climate variability and land cover change impact on flooding and low flows? Hydrol. Processes, 21, 12411247, doi:10.1002/hyp.6669.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Blöschl, G., and Coauthors, 2015: Increasing river floods: Fiction or reality? Wiley Interdiscip. Rev.: Water, 2, 329344, doi:10.1002/wat2.1079.

  • Bouwer, L. M., 2011: Have disaster losses increased due to anthropogenic climate change? Bull. Amer. Meteor. Soc., 92, 3946, doi:10.1175/2010BAMS3092.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Brigode, P., and Coauthors, 2014: Sensitivity analysis of SCHADEX extreme flood estimations to observed hydrometeorological variability. Water Resour. Res., 50, 353370, doi:10.1002/2013WR013687.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Butler, D. R., 1989: The failure of beaver dams and resulting outburst flooding: A geomorphic hazard of the southeastern Piedmont. Geogr. Bull., 31, 2938.

    • Search Google Scholar
    • Export Citation
  • Ceppi, A., G. Ravazzani, A. Salandin, D. Rabuffetti, A. Montani, E. Borgonovo, and M. Mancini, 2013: Effects of temperature on flood forecasting: Analysis of an operative case study in Alpine basins. Nat. Hazards Earth Syst. Sci., 13, 10511062, doi:10.5194/nhess-13-1051-2013.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Condon, L. E., S. Gangopadhyay, and T. Pruitt, 2015: Climate change and non-stationary flood risk for the upper Truckee River basin. Hydrol. Earth Syst. Sci., 19, 159175, doi:10.5194/hess-19-159-2015.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Dahlke, H. E., Z. M. Easton, D. R. Fuka, S. W. Lyon, and T. S. Steenhuis, 2009: Modelling variable source area dynamics in a CEAP watershed. Ecohydrology, 2, 337349, doi:10.1002/eco.58.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • DeGaetano, A. T., 2009: Time-dependent changes in extreme-precipitation return-period amounts in the continental United States. J. Appl. Meteor. Climatol., 48, 20862099, doi:10.1175/2009JAMC2179.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Diffenbaugh, N. S., J. S. Pal, R. J. Trapp, and F. Giorgi, 2005: Fine-scale processes regulate the response of extreme events to global climate change. Proc. Natl. Acad. Sci. USA, 102, 15 77415 778, doi:10.1073/pnas.0506042102.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Easton, Z. M., P. Gérard-Marchant, M. T. Walter, A. M. Petrovic, and T. S. Steenhuis, 2007: Hydrologic assessment of an urban variable source watershed in the northeast United States. Water Resour. Res., 43, W03413, doi:10.1029/2006WR005076.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Elbialy, S., A. Mahmoud, B. Pradhan, and M. Buchroithner, 2014: Application of spaceborne synthetic aperture radar data for extraction of soil moisture and its use in hydrological modelling at Gottleuba catchment, Saxony, Germany. J. Flood Risk Manage., 7, 159175, doi:10.1111/jfr3.12037.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Freudiger, D., I. Kohn, K. Stahl, and M. Weiler, 2014: Large scale analysis of changing frequencies of rain-on-snow events and their impact on floods. Hydrol. Earth Syst. Sci., 18, 26952709, doi:10.5194/hess-18-2695-2014.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Frumhoff, P. C., J. J. McCarthy, J. M. Melillo, S. C. Moser, and D. J. Wuebbles, 2007: Confronting climate change in the U.S. Northeast: Science, impacts, and solutions. Northeast Climate Impacts Assessment Rep., Union of Concerned Scientists, 146 pp. [Available online at http://www.ucsusa.org/sites/default/files/legacy/assets/documents/global_warming/pdf/confronting-climate-change-in-the-u-s-northeast.pdf.]

  • Fuka, D. R., M. T. Walter, J. A. Archibald, T. S. Steenhuis, and Z. M. Easton, 2013: EcoHydRology: A community modeling foundation for eco-hydrology, version 4.9. R package. [Available online at https://cran.r-project.org/web/packages/EcoHydRology/index.html.]

  • Genest, C., and A. C. Favre, 2007: Everything you always wanted to know about copula modeling but were afraid to ask. J. Hydrol. Eng., 12, 347368, doi:10.1061/(ASCE)1084-0699(2007)12:4(347).

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Gersonius, B., R. Ashley, A. Pathirana, and C. Zevenbergen, 2013: Climate change uncertainty: Building flexibility into water and flood risk infrastructure. Climatic Change, 116, 411423, doi:10.1007/s10584-012-0494-5.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Gilroy, K. L., and R. H. McCuen, 2012: A nonstationary flood frequency analysis method to adjust for future climate change and urbanization. J. Hydrol., 414–415, 4048, doi:10.1016/j.jhydrol.2011.10.009.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Grusson, Y., X. Sun, S. Gascoin, S. Sauvage, S. Raghavan, F. Anctil, and J. M. Sáchez-Pérez, 2015: Assessing the capability of the SWAT model to simulate snow, snow melt and streamflow dynamics over an alpine watershed. J. Hydrol., 531, 574588, doi:10.1016/j.jhydrol.2015.10.070.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Haberlandt, U., and I. Radtke, 2014: Hydrological model calibration for derived flood frequency analysis using stochastic rainfall and probability distributions of peak flows. Hydrol. Earth Syst. Sci., 18, 353365, doi:10.5194/hess-18-353-2014.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Harrison, B., and R. Bales, 2015: Skill assessment of water supply outlooks in the Colorado River basin. Hydrology, 2, 112131, doi:10.3390/hydrology2030112.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Hayhoe, K., and Coauthors, 2007: Past and future changes in climate and hydrological indicators in the US Northeast. Climate Dyn., 28, 381407, doi:10.1007/s00382-006-0187-8.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Hayhoe, K., and Coauthors, 2008: Regional climate change projections for the Northeast USA. Mitigation Adapt. Strategies Global Change, 13, 425436, doi:10.1007/s11027-007-9133-2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Hirabayashi, Y., R. Mahendran, S. Koirala, L. Konoshima, D. Yamazaki, S. Watanabe, H. Kim, and S. Kanae, 2013: Global flood risk under climate change. Nat. Climate Change, 3, 816821, doi:10.1038/nclimate1911.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Hirsch, R. M., and K. R. Ryberg, 2012: Has the magnitude of floods across the USA changed with global CO2 levels? Hydrol. Sci. J., 57, 19, doi:10.1080/02626667.2011.621895.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Jabareen, Y., 2013: Planning the resilient city: Concepts and strategies for coping with climate change and environmental risk. Cities, 31, 220229, doi:10.1016/j.cities.2012.05.004.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Jörg-Hess, S., N. Griessinger, and M. Zappa, 2015: Probabilistic forecasts of snow water equivalent and runoff in mountainous areas. J. Hydrometeor., 16, 21692186, doi:10.1175/JHM-D-14-0193.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Knighton, J., and M. Walter, 2016: Critical rainfall statistics for predicting watershed flood responses: Rethinking the design storm concept. Hydrol. Processes, 30, 37883803, doi:10.1002/hyp.10888.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Koplin, N., B. Schadler, D. Viviroli, and R. Weingartner, 2014: Seasonality and magnitude of floods in Switzerland under future climate change. Hydrol. Processes, 28, 25672578, doi:10.1002/hyp.9757.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Kundzewicz, Z. W., and Coauthors, 2014: Flood risk and climate change: Global and regional perspectives. Hydrol. Sci. J., 59, 128, doi:10.1080/02626667.2013.857411.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Kunkel, K. E., and Coauthors, 2013: Monitoring and understanding trends in extreme storms: State of knowledge. Bull. Amer. Meteor. Soc., 94, 499514, doi:10.1175/BAMS-D-11-00262.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Lawrence, D., E. Paquet, J. Gailhard, and A. K. Fleig, 2014: Stochastic semi-continuous simulation for extreme flood estimation in catchments with combined rainfall–snowmelt flood regimes. Nat. Hazards Earth Syst. Sci., 14, 12831298, doi:10.5194/nhess-14-1283-2014.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Li, J., P. Feng, and F. Chen, 2014: Effects of land use change on flood characteristics in mountainous area of Daqinghe watershed, China. Nat. Hazards, 70, 593607, doi:10.1007/s11069-013-0830-8.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Lo, M.-H., and J. S. Famiglietti, 2010: Effect of water table dynamics on land surface hydrologic memory. J. Geophys. Res., 115, D22118, doi:10.1029/2010JD014191.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Mahanama, S., B. Livneh, R. Koster, D. Lettenmaier, and R. Reichle, 2012: Soil moisture, snow, and seasonal streamflow forecasts in the United States. J. Hydrometeor., 13, 189203, doi:10.1175/JHM-D-11-046.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Mallakpour, I., and G. Villarini, 2015: The changing nature of flooding across the central United States. Nat. Climate Change, 5, 250254, doi:10.1038/nclimate2516.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Massari, C., L. Brocca, S. Barbetta, C. Papathanasiou, M. Mimikou, and T. Moramarco, 2014a: Using globally available soil moisture indicators for flood modelling in Mediterranean catchments. Hydrol. Earth Syst. Sci., 18, 839853, doi:10.5194/hess-18-839-2014.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Massari, C., L. Brocca, T. Moramarco, Y. Tramblay, and J. F. D. Lescot, 2014b: Potential of soil moisture observations in flood modelling: Estimating initial conditions and correcting rainfall. Adv. Water Resour., 74, 4453, doi:10.1016/j.advwatres.2014.08.004.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Massari, C., A. Tarpanelli, L. Brocca, and T. Moramarco, 2015: Assimilating satellite soil moisture into rainfall–runoff modelling: Towards a systematic study. Geophysical Research Abstracts, Vol. 17, Abstract EGU2015-1780. [Available online at http://meetingorganizer.copernicus.org/EGU2015/EGU2015-1780.pdf.]

  • Moriasi, D. N., J. G. Arnold, M. W. Van Liew, R. L. Bingner, R. D. Harmel, and T. L. Veith, 2007: Model evaluation guidelines for systematic quantification of accuracy in watershed simulations. Trans. ASABE, 50, 885900, doi:10.13031/2013.23153.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Nash, J. E., and J. V. Sutcliffe, 1970: River flow forecasting through conceptual models part I—A discussion of principles. J. Hydrol., 10, 282290, doi:10.1016/0022-1694(70)90255-6.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Nied, M., Y. Hundecha, and B. Merz, 2013: Flood-initiating catchment conditions: A spatio-temporal analysis of large-scale soil moisture patterns in the Elbe River basin. Hydrol. Earth Syst. Sci., 17, 14011414, doi:10.5194/hess-17-1401-2013.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • NOAA/HDSC, 2016: Precipitation Frequency Data Server (PFDS). Accessed 28 December 2016. [Available online at http://hdsc.nws.noaa.gov/hdsc/pfds/index.html.]

  • NOAA/NCEI, 2015: Quality controlled local climatological data. Accessed 28 December 2016. [Available online at http://www.ncdc.noaa.gov/data-access/land-based-station-data/land-based-datasets.]

  • Norbiato, D., M. Borga, S. Degli Esposti, E. Gaume, and S. Anquetin, 2008: Flash flood warning based on rainfall thresholds and soil moisture conditions: An assessment for gauged and ungauged basins. J. Hydrol., 362, 274290, doi:10.1016/j.jhydrol.2008.08.023.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • NRCC, 2015: NRCC hourly precipitation database. Accessed 28 December 2016. [Available online at http://www.nrcc.cornell.edu/services/access/access.html.]

  • NRCS, 2015: Web Soil Survey. Accessed 11 November 2016. [Available online at http://websoilsurvey.sc.egov.usda.gov/App/HomePage.htm.]

  • Obeysekera, J., and J. D. Salas, 2014: Quantifying the uncertainty of design floods under nonstationary conditions. J. Hydrol. Eng., 19, 14381446, doi:10.1061/(ASCE)HE.1943-5584.0000931.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Palecki, M. A., J. R. Angel, and S. E. Hollinger, 2005: Storm precipitation in the United States. Part I: Meteorological characteristics. J. Appl. Meteor., 44, 933946, doi:10.1175/JAM2243.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Paquet, E., F. Garavaglia, R. Garçon, and J. Gailhard, 2013: The SCHADEX method: A semi-continuous rainfall–runoff simulation for extreme flood estimation. J. Hydrol., 495, 2337, doi:10.1016/j.jhydrol.2013.04.045.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Paschalis, A., S. Fatichi, P. Molnar, S. Rimkus, and P. Burlando, 2014: On the effects of small scale space–time variability of rainfall on basin flood response. J. Hydrol., 514, 313327, doi:10.1016/j.jhydrol.2014.04.014.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Perju, E. R., D. Balin, S. Lane, and L. Zaharia, 2013: Changing flood magnitude and frequency in snow-melt dominated catchments: The case of the Bucegi Mountains, in the Romanian Carpathian region. Geophysical Research Abstracts, Vol. 15, Abstract EGU2013-10110. [Available online at http://meetingorganizer.copernicus.org/EGU2013/EGU2013-10110.pdf.]

  • Ravazzani, G., M. Mancini, I. Giudici, and P. Amadio, 2007: Effects of soil moisture parameterization on a real-time flood forecasting system based on rainfall thresholds. IAHS Publ., 313, 407–416.

  • Rossman, L. A., 2010: Storm water management model user’s manual, version 5.0. Rep. EPA/600/R-05/040, Environmental Protection Agency, 285 pp.

  • Salas, J. and Obeysekera, J. 2014: Revisiting the concepts of return period and risk for nonstationary hydrologic extreme events. J. Hydrol. Eng., 19, 554568, doi:10.1061/(ASCE)HE.1943-5584.0000820.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Schiermeier, Q., 2011: Increased flood risk linked to global warming. Nature, 470, 316316, doi:10.1038/470316a.

  • SCS, 1986: Urban hydrology for small watersheds. USDA TR-55, 164 pp. [Available online at https://www.nrcs.usda.gov/Internet/FSE_DOCUMENTS/stelprdb1044171.pdf.]

  • Seidou, O., A. Ramsay, and I. Nistor, 2012: Climate change impacts on extreme floods I: Combining imperfect deterministic simulations and non-stationary frequency analysis. Nat. Hazards, 61, 647659, doi:10.1007/s11069-011-0052-x.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Serinaldi, F., 2015: Dismissing return periods! Stochastic Environ. Res. Risk Assess., 29, 11791189, doi:10.1007/s00477-014-0916-1.

  • Smith, B. K., J. A. Smith, M. L. Baeck, and A. J. Miller, 2015: Exploring storage and runoff generation processes for urban flooding through a physically based watershed model. Water Resour. Res., 51, 15521569, doi:10.1002/2014WR016085.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Stedinger, J. R., and V. W. Griffis, 2011: Getting from here to where? Flood frequency analysis and climate. J. Amer. Water Resour. Assoc., 47, 506513, doi:10.1111/j.1752-1688.2011.00545.x.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Surfleet, C., and D. Tullos, 2013: Variability in effect of climate change on rain-on-snow peak flow events in a temperate climate. J. Hydrol., 479, 2434, doi:10.1016/j.jhydrol.2012.11.021.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Taylor, K. E., R. Stouffer, and G. Meehl, 2012: An overview of CMIP5 and the experiment design. Bull. Amer. Meteor. Soc., 93, 485498, doi:10.1175/BAMS-D-11-00094.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Tolson, B. A., and C. A. Shoemaker, 2007: Dynamically dimensioned search algorithm for computationally efficient watershed model calibration. Water Resour. Res., 43, W01413, doi:10.1029/2005WR004723.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Tramblay, Y., L. Neppel, J. Carreau, and K. Najib, 2013: Non-stationary frequency analysis of heavy rainfall events in southern France. Hydrol. Sci. J., 58, 280294, doi:10.1080/02626667.2012.754988.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Tramblay, Y., E. Amoussou, W. Dorigo, and G. Mahé, 2014: Flood risk under future climate in data sparse regions: Linking extreme value models and flood generating processes. J. Hydrol., 519, 549558, doi:10.1016/j.jhydrol.2014.07.052.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Trenberth, K. E., 2011: Changes in precipitation with climate change. Climate Res., 47, 123138, doi:10.3354/cr00953.

  • USGS, 2015: USGS 04234000 Fall Creek near Ithaca, NY. Accessed 11 November 2016. [Available online at http://waterdata.usgs.gov/nwis/uv?04234000.]

  • USGS, 2016: National elevation dataset. Accessed 11 November 2016. [Available online at http://nationalmap.gov/elevation.html.]

  • Walter, M., P. Gérard-Marchant, T. Steenhuis, and M. Walter, 2005: Closure to “Simple estimation of prevalence of Hortonian flow in New York City watersheds” by M. Todd Walter, Vishal K. Mehta, Alexis M. Marrone, Jan Boll, Pierre Gérard-Marchant, Tammo S. Steenhuis, and Michael F. Walter. J. Hydrol. Eng., 10, 168169, doi:10.1061/(ASCE)1084-0699(2005)10:2(169).

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Wehner, M. F., 2013: Very extreme seasonal precipitation in the NARCCAP ensemble: Model performance and projections. Climate Dyn., 40, 5980, doi:10.1007/s00382-012-1393-1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Weijs, S. V., N. van de Giesen, and M. B. Parlange, 2013a: Data compression to define information content of hydrological time series. Hydrol. Earth Syst. Sci., 17, 31713187, doi:10.5194/hess-17-3171-2013.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Weijs, S. V., N. van de Giesen, and M. B. Parlange, 2013b: HydroZIP: How hydrological knowledge can be used to improve compression of hydrological data. Entropy, 15, 12891310, doi:10.3390/e15041289.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Westra, S., I. Varley, P. Jordan, R. Nathan, A. Ladson, A. Sharma, and P. Hill, 2010: Addressing climatic non-stationarity in the assessment of flood risk. Aust. J. Water Resour., 14, 116.

    • Search Google Scholar
    • Export Citation
  • Wood, A. W., and J. C. Schaake, 2008: Correcting errors in streamflow forecast ensemble mean and spread. J. Hydrometeor., 9, 132148, doi:10.1175/2007JHM862.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Wood, A. W., T. Hopson, A. Newman, L. Brekke, J. Arnold, and M. Clark, 2016: Quantifying streamflow forecast skill elasticity to initial condition and climate prediction skill. J. Hydrometeor., 17, 651668, doi:10.1175/JHM-D-14-0213.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Wyżga, B., J. Zawiejska, and A. Radecki-Pawlik, 2016: Impact of channel incision on the hydraulics of flood flows: Examples from Polish Carpathian rivers. Geomorphology, 272, 1020, doi:10.1016/j.geomorph.2015.05.017.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Ye, H., E. J. Fetzer, A. Behrangi, S. Wong, B. H. Lambrigtsen, C. Y. Wang, J. Cohen, and B. Gamelin, 2016: Increasing daily precipitation intensity associated with warmer air temperatures over Northern Eurasia. J. Climate, 29, 623636, doi:10.1175/JCLI-D-14-00771.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Yossef, N. C., H. Winsemius, A. Weerts, R. Beek, and M. F. Bierkens, 2013: Skill of a global seasonal streamflow forecasting system, relative roles of initial conditions and meteorological forcing. Water Resour. Res., 49, 46874699, doi:10.1002/wrcr.20350.

    • Crossref
    • Search Google Scholar
    • Export Citation
Save