Tracks and Dates of the selected tropical cyclones in 2005, 2007 and 2008. 

Contexts in source publication

Context 1

... on the track and intensity forecasts for the eight hurricanes shown in Figure 1 is given in subsection 4.1. Analysis on the forecasts of rainfall, soil moisture and runoff for hurricane Katrina is given in subsection 4.2. The results of streamflow forecast for hurricane Katrina is given in subsection 4.3. A brief summary is given in subsection 4.4 4.1 Track and Intensity Comparison for the Eight Hurricanes To look at the effects of the replacement of the LSM on hurricane track and intensity forecasts, 28 runs were conducted for the eight hurricanes. The HWRF model is initialized with GFS forecasting in all runs. As shown in Figure 3, the effects of the replacement of the LSM in the HWRF on hurricane track and intensity forecasts are very small for the eight cases shown in Figure 1. This suggests that the SLAB LSM can be replaced by Noah LSM in the HWRF. 4.2 Rainfall, Soil Moisture and Runoff for Katrina Two runs were conducted for Katrina using GFS and NAM soil moisture as initials, respectively to look at the initial soil moisture effects on the forecasted rainfall, runoff and streamflow. The run using GFS soil moisture as the initial is referred as HWRF WET run while the run using NAM soil moisture as the initial is referred as HWRF DRY run later in the text for convenience. 12-hour accumulated rainfall for both of the observations and simulations are shown in Figure 4. At 00Z on 29 August 2005, there were some widely spread light rainfall over the eastern CONUS. Both WET and DRY run captured most of the coastal rainfall, but missed the rainfall inland. At 00z on 30 August 2005, the center of simulated rainfall at the coast of Louisiana matches well with the observations. However, the model missed the rainfall over the Ohio valley. At 00z on 31 August 2005, the simulated rainfall center is located to the southwest of the observed, and the model missed most of the rainfall over the northeast of US. There is little difference between the wet run and dry run, indicating that the initial soil moisture had a very small effect on rainfall prediction. However, the effect of initial soil moisture on soil moisture prediction is significant as shown in Figure 5 and 6. At 00Z on 29, 30 and 31 August 2005, soil moisture in layer ...

Context 2

... on the track and intensity forecasts for the eight hurricanes shown in Figure 1 is given in subsection 4.1. Analysis on the forecasts of rainfall, soil moisture and runoff for hurricane Katrina is given in subsection 4.2. The results of streamflow forecast for hurricane Katrina is given in subsection 4.3. A brief summary is given in subsection 4.4 4.1 Track and Intensity Comparison for the Eight Hurricanes To look at the effects of the replacement of the LSM on hurricane track and intensity forecasts, 28 runs were conducted for the eight hurricanes. The HWRF model is initialized with GFS forecasting in all runs. As shown in Figure 3, the effects of the replacement of the LSM in the HWRF on hurricane track and intensity forecasts are very small for the eight cases shown in Figure 1. This suggests that the SLAB LSM can be replaced by Noah LSM in the HWRF. 4.2 Rainfall, Soil Moisture and Runoff for Katrina Two runs were conducted for Katrina using GFS and NAM soil moisture as initials, respectively to look at the initial soil moisture effects on the forecasted rainfall, runoff and streamflow. The run using GFS soil moisture as the initial is referred as HWRF WET run while the run using NAM soil moisture as the initial is referred as HWRF DRY run later in the text for convenience. 12-hour accumulated rainfall for both of the observations and simulations are shown in Figure 4. At 00Z on 29 August 2005, there were some widely spread light rainfall over the eastern CONUS. Both WET and DRY run captured most of the coastal rainfall, but missed the rainfall inland. At 00z on 30 August 2005, the center of simulated rainfall at the coast of Louisiana matches well with the observations. However, the model missed the rainfall over the Ohio valley. At 00z on 31 August 2005, the simulated rainfall center is located to the southwest of the observed, and the model missed most of the rainfall over the northeast of US. There is little difference between the wet run and dry run, indicating that the initial soil moisture had a very small effect on rainfall prediction. However, the effect of initial soil moisture on soil moisture prediction is significant as shown in Figure 5 and 6. At 00Z on 29, 30 and 31 August 2005, soil moisture in layer ...

Context 3

... the grid cell through (at least) one of eight directions, the runoff transport process is linear and time invariant, and the causality and the impulse response functions are nonnegative (Lohmann et al., 2004). With use of the Noah LSM, the HWRF prediction of runoff can be used to drive the EMC streamflow routing scheme to produce HWRF forecasts of streamflow and river flow. The replacement of the Slab LSM with the Noah LSM can enhance the HWRF forecast capability. Tests and evaluations were conducted to ensure that the replacement of the LSM will not degrade forecasts of hurricane track and intensity. Eight tropical cyclones (Figure 1) were selected based on their tracks in this study. KATRINA, DENNIS, RITA, IKE, and GUSTAV made their way to the southern US via the Gulf of Mexico, HANNA moved along the east coast of Atlantic Ocean, DEAN made its way to Mexico while FAY stopped over several islands before making landfall over Florida. Two different runs for each case were conducted with the Slab LSM and Noah LSM, respectively. These experiments are summarized in Table 1. Table 1. Summary of experiments to look at the LSM effects on hurricane track and intensity forecasts Runs Physics Options Over Land LSM Surface Layer Scheme PBL Scheme Control Run (N883) Slab GFDL GFS Test Run (N893) Noah GFDL GFS Among the selected hurricanes, the HWRF model did the best job for Katrina in terms of track, intensity and rainfall forecast. As such, Katrina was selected to test the river routing scheme. The current operations of the HWRF forecast use GFS outputs as initial fields, including soil moisture. GFS soil moisture is driven by GDAS predicted precipitation, plus nudging towards a monthly soil moisture climatology. The soil moisture for NLDAS and NAM are driven by observed precipitation. Therefore, NAM soil moisture is more realistic than GFS soil moisture, for example, the GFS soil ...