River flow forecasting using soft computing and remote sensing data (case study: HableRud Basin)

Runoff estimation in Dry lands is very complicated due to numerous factors that impact
on runoff generation and so temporal and spatial scale conflicts. This condition will be
more conflict because of population growth and land use change. Also, climate change
can influence outflow change due to changing climatic variables. On the other hand out
climatic signals can impact on climate condition in each region. River flow forecasting
can cause optimal water resources management and better economical justify of water
resources projects. In this study data sensitivity analysis was performed on local and out
climatic data namely
Teleconnection indexes. Four soft computing methods of Generalized Feed forward
Neural Networks (JFFN), Jordan-Elman Neural Networks(JEN), Time Lag Neural
Networks(TLNN) and Co Adaptive Neural Fuzzy Interface Systems(CANFIS) used in
Four sub-basins in two time periods of 30 and 11 years. Satellite maps of MODIS
Landsat used for calculating snow cover. Sensitivity analysis of Teleconnection indexes
performed to clear its impact on the river flow. Outputs of general circulation models of
HADCM3, IPCM4 and BCM2 under scenarios of A1B, A2 and B1 downscaled by
LARS-WG statistical model during 3 time steps of 2011-2030, 2046-2065 and 2080-
2099. IHACRES model was used for future runoff projection. Results indicated that
snow cover area can decrease modeling process error especially using models with time
lag algorithm. Soft computing models can decrease error for river flow forecasting,
although better result can gain with shorter time period and snow cover parameter.
Teleconnection indexes can improve river flow forecasting. Also, projected results
showed that river flow volume will be diminished till the end of century about 11.1
percent and it means a 43.5 million cubic meter decrease in the outflow of basin 4.these
results can effective in optimal water resources management in the study area and
increase our knowledge for river flow trend forecasting.

River flow forecasting using soft computing and remote sensing data (case study: HableRud Basin) Keywords:

Soft computing , river flow , remote sensing , snow cover , Teleconnection indexes