Application of regression in seasonal flow forecasting for Upper Indus Basin of Pakistan
- PDF / 3,937,240 Bytes
- 11 Pages / 595.276 x 790.866 pts Page_size
- 9 Downloads / 193 Views
ORIGINAL PAPER
Application of regression in seasonal flow forecasting for Upper Indus Basin of Pakistan Muhammad Umar 1 Received: 14 April 2020 / Accepted: 15 September 2020 # Saudi Society for Geosciences 2020
Abstract Water managers in Pakistan need timely and accurate seasonal flow forecasts like in several areas of the world to allocate water for various kinds of water usages like canal operations for irrigation, reservoir operations, and strategies to respond to extreme cases. Various hydrological models are deployed that include University of British Columbia Watershed Model (UBCWM) and snowmelt runoff model (SRM) to seasonal streamflow forecasts in Pakistan. Here, I assess the approach that employed snow water equivalent (SWE), temperature (T), precipitation (P), and February month flow (Feb) to forecasts Kharif (April– September) season with no need for intensive hydrological modeling in a skillful way for Upper Indus Basin (UIB) at Tarbela Dam. For Indus, I compare this approach results with Indus River System Authority (IRSA), UBCWM, and SRM. This approach uses multiple linear regression (MLR) to develop regression function for forecast seasonal flow volume. This regression approach provides much skillful flow forecasts that is consistent in uncertainty spread for the Indus. The regression forecast accuracy with a mean absolute percentage error (MAPE) is 7.95% with regard to statistical approach used by IRSA, UBCWM, and SRM as 10.25%, 11.05%, and 8.91%, respectively. This approach even allows improvement of 1% (volume ~ 0.6 km3) and is very simple to apply with requirement of four input data that are easy to procure or download and process. With this information in hand, a regression function can generate seasonal forecast for the authorities in Pakistan. Keywords Streamflow . Forecasting . Snow water equivalent . Precipitation . Temperature . Multiple linear regression
Introduction Water managers in Pakistan need timely and accurate seasonal flow forecasts like in several areas of the world to allocate water for various kinds of water usages (Wood et al. 2016). Snow and glacier mountains are considered to be the water towers in the world. Nearly half of the world get freshwater from these largest reservoirs to fulfil their needs for domestic, irrigation, hydropower, and industrial usages (Viviroli et al. 2007). The Indus River is considered to be the bread basket on which it depends on all socioeconomic development of Pakistan (Cook et al. 2013; Rieu-Clarke 2015). Pakistan as an agrarian economy, agriculture part in water usage is nearly
Responsible Editor: Broder J. Merkel * Muhammad Umar [email protected] 1
National Engineering Services Pakistan (Pvt) Limited (NESPAK), Lahore, Pakistan
97%, which is 27% higher than the global average (Akram Qazi et al. 2009). The provincial water shares are decided by Indus River System Authority (IRSA), Pakistan, according to the 1991 Water Apportionment Accord (WAA) (Basharat et al. 2014) (Anwar and Bhatti 2018), and seasonal allocations of water to various
Data Loading...
