Table 2 Improved downscaling in RectifHydPlus attributable to new proxy data and bespoke spill procedure.

From: Forty-year hydropower generation reanalysis for Conterminous United States

 

RectifHyd median score

RectifHydPlus median score

% of plant-years improved with RectifHydPlus

Proxy A (2,405 dam-years)

 NSE

0.89

0.96

77%

 KGE

0.81

0.92

76%

 R squared

0.92

0.97

72%

Proxy B* (1,698 dam-years)

 NSE

−0.01

0.78

86%

 KGE

0.23

0.75

76%

 R squared

0.31

0.84

80%

Proxy C* (1,169 dam-years)

 NSE

0.02

0.60

84%

 KGE

0.42

0.70

78%

 R squared

0.44

0.69

76%

Proxy D* (2,446 dam-years)

 NSE

0.19

0.57

73%

 KGE

0.44

0.67

70%

 R squared

0.43

0.69

69%

Proxy E (1,851 dam-years)

 NSE

−0.03

0.19

73%

 KGE

0.27

0.41

62%

 R squared

0.33

0.37

60%

  1. Performance scores computed for 510 plants and 21 years (10,700 plant-years). Performance scores are computed using observed monthly data available 1980 – 2000. *Data are split according to the proxy data used to downscale annual generation in RectifHydPlus (not RectifHyd, which lacks proxies B, C, and D)—meaning results reported for RectifHyd in proxies B, C, and D reflect performance in the absence of those new proxies.