EU FUNDED FP7 PROJECT FLOODSAT (277183)

EU flag http://www.metu.edu.tr/%7Eyilmazk/html/img16.jpg                                                                                                                                                        http://www.metu.edu.tr/%7Eyilmazk/html/img15.jpg

 

The overarching objective of the FLOODSAT project is to advance the utility of satellite-based precipitation estimates for hydrologic modeling, specifically for flood monitoring. The main objectives include evaluation of the satellite-based precipitation (SBP) products over the Western Black Sea Region in Turkey using the rain gauge network (Figure 1), devising a bias-adjustment algorithm for the SBP products and implementation of a hydrologic model in a selected watershed. An improved methodology for calibration and evaluation of the hydrological model utilizing streamflow observations measured at the outlet and interior points will be devised. The performance of the hydrologic model driven by satellite-based and ground-based observations will provide further insights into the value of existing and adjusted satellite-based precipitation estimates for basin-scale streamflow simulations.

Figure 1. Western Black Sea Region, Turkey. Meteorological stations within windward side of mountains are marked by a “dot marker” and those within leeward side are marked by a “triangle marker”.

The Satellite-based precipitation (SBP) products utilized in this study include The Tropical Rainfall Measuring Mission (TRMM) Multi-satellite Precipitation Analysis (Version 7) including real time (TMPA-7RT) and gauge adjusted (TMPA-7A) products; Climate Prediction Center Morphing Technique (CMORPH) and the Multi Sensor Precipitation Estimate (MPE) of the European Organization for the Exploitation of Meteorological Satellites (EUMETSAT). The study area is characterized by a complex topography, marked by Northeast-Southwest aligned mountain ranges running parallel to the shoreline. Hence orographic precipitation is very characteristic at the windward (North) side of the mountains. A rain gauge-based gridded precipitation product was constructed considering the “physiographic similarity” concept in which a locally weighted linear rainfall-elevation relationship was used for precipitation estimation at 0.05o grids. The weights of the rain gauges were established using a combination of a set of physiographic descriptors including distance, elevation, coastal proximity, facet and effective terrain height.

Evaluation of the SBP products indicated that the orographic controls on the precipitation influences the performance of the SBP products (Figure 2). Evaluated SBP products generally had difficulty in representing the precipitation gradient normal to the orography. TMPA-7RT, TMPA-7A and MPE products underestimated precipitation along the windward region and overestimated the precipitation on the leeward region, more significantly during the cold season. CMORPH product underestimated the precipitation on both windward and leeward regions regardless of the season. Further investigation of the datasets used in the development of these SBP products revealed that, although both infrared (IR) and microwave (MW) datasets contain potential problems, inability of MW sensors to detect precipitation especially in cold season was the main challenge over this region with complex topography.

 

Figure 2. (a) Topographic elevations, (b) annual precipitation for each SBP product and rain gauge along cross section line 1.

A new bias adjustment methodology (BAPS) for the satellite-based precipitation products has been devised based on the “physiographic similarity” concept. In the proposed procedure, SBP estimates are adjusted based on a weighted scheme using rain gauges. Rain gauge weights were calculated based on physiographic similarity concept which is more suited to the complex terrain as opposed to the commonly used proximity concept.  Comparison of SBP products before and after bias adjustment with BAPS algorithm indicated that the adjustment procedure successfully corrected for the precipitation detection problems. The bias adjustment procedure resulted in significant improvements in the performance of the CMORPH and MPE products with less significant improvements in TMPA-7RT product showing high spatial heterogeneity in precipitation estimates (Figure 3). Comparison of the BAPS algorithm with the baseline Inverse Distance Weighted Algorithm (IDW) using two independent stations showed that the BAPS algorithm generally provided better statistics.

Figure 3. Box-plots showing seasonal monthly statistical results calculated between RGP grids in windward (Yellow box) and leeward (White box) regions and co-located SBP products before bias adjustment (a-c) and after bias adjustment (d-f). Note that TMPA-7A already includes rain gauge information and hence has not been adjusted.

 

A coupled distributed hydrological model, MIKE-SHE/MIKE-11 was implemented for the Karabük Basin located within the study area (Western Black Sea Basin of Turkey). Streamflow is observed at two stations in the Karabük Basin, at the outlet and at an interior location marking the outlet of the Araç Subbasin (Figure 4). A novel model calibration strategy has been devised that builds on the signature measure approach. In the proposed strategy, model parameters controlling different hydrological processes are grouped and constrained in a hierarchical, step-wise manner using signature measures sensitized to extract streamflow information relevant to the focused hydrological process in each step. The calibration procedure starts with random sampling of parameter sets from the feasible parameter space and calculation of signature measures extracted from streamflow observations.  The procedure follows a bottom-up hierarchical order in the soil column where parameters controlling the groundwater component (baseflow) are constrained first using the selected low flow signature measure. In a hierarchical, step-wise procedure, parameters controlling interflow, unsaturated flow, overland flow and channel flow are constrained using signature measures selected for the focused hydrological process in each step. In the final stage of the hierarchical constraining procedure, a parameter set that provides the best performance in terms of all signature measures was selected and tested within the evaluation period. The proposed calibration strategy was first utilized in the interior  Arac  sub-basin  using  the   RGP  dataset.   The MIKE-SHE/MIKE-11   model  with  the parameter values calibrated using the RGP dataset was then utilized to evaluate the value of satellite-based precipitation products (before and after bias adjustment) for streamflow simulation. It was found that the errors in the satellite-based precipitation products directly translates into the simulated flows, indicated by overestimation of flows when the model was driven by TMPA products and underestimation of flows when the model was driven by CMORPH product. Bias adjustment using the proposed BAPS algorithm improved the flow simulation performance significantly. It also was found that the model performance did not improve significantly when the model parameters were re-calibrated using satellite-based products instead of using RGP-calibrated parameters (Figures 5 & 6). These results indicated that bias adjustment of satellite-based rainfall products using relatively sparse raingauge network will improve the hydrologic model performance in simulating flood events in the study area.

 

Figure 4. Karabük and Araç Basins in which the MIKE-SHE/MIKE-11 model was implemented.

 

Figure 5. Simulated and observed (a) Flow duration curve and (b) hydrograph during calibration period and (c) hydrograph during evaluation period for RGP-calibrated model driven by TMPA-7RT BAPS in Arac Basin.

Figure 6. Simulated and observed (a) FDC and (b) hydrograph during calibration period and (c) hydrograph during evaluation period for TMPA-7RT BAPS -calibrated model driven by TMPA-7RT BAPS product in Arac Basin.

 

In an effort to check the model parameter transferability for the ungauged basins within the study area, the parameter set calibrated for the Arac Basin was transferred to and evaluated for the downstream Karabuk Basin. It was found that the model with the transferred parameters are able to adequately capture the high streamflow events.  

 

At the international level, the project resulted in two important advancement beyond the state of the art. First, a bias adjustment algorithm for satellite-based precipitation estimates (BAPS) has been devised based on the ‘Physiographic similarity’ concept which is well-suited to regions with complex topography with relatively sparse raingauge network. Bias adjustment algorithms in the literature generally rely on the proximity concept and is not suited to topographically complex regions where topographic features significantly impact the precipitation patterns at short distances. Second, a novel model calibration strategy has been devised that builds on the signature measure approach. In the proposed strategy, model parameters controlling different hydrological processes are grouped and constrained in a hierarchical, step-wise manner using signature measures sensitized to extract streamflow information relevant to the focused hydrological process in each step. The calibration procedure starts with random sampling of parameter sets from the feasible parameter space and calculation of signature measures extracted from streamflow observations.  The procedure follows a bottom-up hierarchical order in the soil column where parameters controlling the groundwater component (baseflow), interflow, unsaturated flow, overland flow and channel flow are constrained sequentially using signature measures selected for the focused hydrological process in each step. The use of signature measures, and parameter constraining based on sensitivity analysis are the main advancements beyond the current state of the art in the field of parameter estimation.

 

Project Duration: April 2011-March 2014

 

Publications:

Derin, Y. and Yilmaz, K.K., 2014: Evaluation of Multiple Satellite-Based Precipitation Products over Complex Topography, Journal of Hydrometeorology, 15 (4), 1498-1516. doi:10.1175/JHM-D-13-0191.1

Derin, Y. and Yilmaz, K.K., (in preparation): Bias Adjustment Methodology for Satellite-Based Precipitation Products Over Complex Terrain Based on the concept of Physiographic Similarity. (Draft manuscript attached as Annex)

Derin, Y. and Yilmaz, K.K., (in preparation): A process oriented sequential constraining approach to model evaluation.

Derin, Y. and Yilmaz, K.K., (in preparation): The utility of satellite-based precipitation products for hydrologic modeling: Is bias adjustment necessary?

Scientific Conferences - Oral:

Yilmaz K. K. and Derin, Y., (2014). Advancement of Satellite-based Rainfall Applications for Hydrologic Modeling in Topographically Complex Regions, European Geosciences Union General Assembly, Vienna, Austria, 27 April – 2 May 2014, Geophysical Research Abstracts Vol. 16, EGU2014-13462, 2014

Derin, Y., Yilmaz, K. K., Physically-Based Integrated Hydrologic Modeling in Araç Basin, 67th Geological Congress of Turkey, Ankara, Turkey, 14-18 April 2014.

Derin, Y., Yilmaz, K. K., Evaluation and Bias Adjustment of Multiple Satellite-Based Rainfall Products over Complex Terrain, 67th Geological Congress of Turkey, Ankara, Turkey, 14-18 April 2014.

Scientific Conferences - Poster:

Yilmaz K.K., and Derin, Y., Advancement of Satellite-based Rainfall Applications for Hydrologic Modeling in Topographically Complex Regions, In American Geophysical Union Fall Meeting, San Francisco, USA, 9 – 13 December 2013.

Yilmaz, K. K., Derin, Y., Evaluation and Bias Adjustment of Multiple Satellite-Based Rainfall Products over Complex Terrain, Geology and Remote Sensing Symposium, 13-14 March 2014, Akdeniz University, Antalya, Turkey (in Turkish).

Derin, Y. and Yilmaz K.K., Evaluation of Multiple Satellite-based Rainfall Products over a Topographically Complex Watershed, European Geosciences Union General Assembly, Vienna, Austria, 22 – 27 April 2012.

Derin, Y. and Yilmaz K.K., Evaluation and Bias Adjustment of Multiple Satellite-based Precipitation Products over Complex Terrain, European Geosciences Union General Assembly, Vienna, Austria, 7 – 12 April 2013.

Awards:

Ms. Derin (graduate student funded by FLOODSAT Project) received the prestigious “Hydrological Sciences Outstanding Student Poster Award” at European Geosciences Union General Assembly held in April 2013.

Ms. Derin interviewed by GEOLOG, The Official Blog of the European Geosicences Union. http://geolog.egu.eu/2013/11/13/geotalk-yagmur-derin-on-posters-and-precipitation/