Overview

Six models are used to forecast each major age class of sockeye salmon in the major rivers of Bristol Bay. Five of these models are linear regressions and the sixth is a 5-year running average of the age class under consideration.

Forecast models

This document covers output from five linnear regression models run against four major age classes in this river. Three regressions are variations of the sibling relationship with the age being forecast (Table 1). The first regression (sibling) uses observed forecast age and the brood year younger sibling return. The second regression (ln.sibling) uses a natural log transformation of both the forecasted age and the sibling return. The third regression (lnY.sibling) uses the sibling age return and a natural log transformation of the forecasted age. Two linnear regression models based on forecasted age return and total spawning escapment. The first of these (lnS.lnR) is the natural log of the forecasted age class regressed against the natural log of the spawning escapement. The second forecast method that relies on the relationship between the forecasted age and the spawning escapement is a linnear transformation of the Ricker relationship.

Table 1. Forecast ages and the sibling age class used for forecasting.

Forecast Age Sibling
Age.12 Age.11
Age.22 Age.21
Age.13 Age.12
Age.23 Age.22

Output

Each model output consists of a table and two graphs. The table shows the most recent 10 years of observed returns for the age being forecast (‘Observed Return’) and the relevant sibling return (‘Sibling Return’) shown in bold black text. Brood and return years are shown in blue text. Two regression model statistics (R-squared and P-value of the slope) are shown in red. Model forecasted returns are shown in bold black text with a yellow highlight. Error measurements are shown in green text. Error measurements shown are absolute deviation (AD), absolute percent error (APE), percent error (PE) and arc tan percent error (AAPE).

After the table summarizing the model output for the most recenet ten years, two graphs are shown for each of the linnear regression models. The first graph is a scatter plot of the data with a linnear regression superimposed. The second graph is a plot of the residuals of the regression analysis.

Absolute deviation (AD) is defined as

\[\ AD = O - F\] where

O = Observed Return

F = Forecasted Return

Percent error (PE) is defined as

\[\ PE = (F - O)/O\] Absolute Percent error (APE) is defined as

\[\ APE = |(F - O)/O|\]

Arctangent percent error (AAPE) is defined (Kim and Kim, 2016) as \[\ AAPE = arctan(|((O - F)/O)|)\]

At the end of the model output section for each age class, a performance table is presented that contains the mean absolute deviation (MAD), neam absolute percent error (MAPE), mean percent error (MPE) and mean arctanget percent error (MAAPE). These performance metrics are calculated and presented across the most recent three and five years. Color bars show the relative size of each metric.

age.12

Below are model outputs using data from 1995

lnS.lnR

Return Yr Brood Yr Observed Return Spawner Return R Sq P-val Forecasted Return AD APE PE AAPE
2013 2009 460,969 970,818 0.153 0.092 221,743 239,226 1.08 -1.08 0.48
2014 2010 735,100 1,187,730 0.151 0.085 305,491 429,610 1.41 -1.41 0.53
2015 2011 3,087,468 883,794 0.149 0.077 220,857 2,866,610 12.98 -12.98 0.75
2016 2012 1,123,765 861,747 0.106 0.109 262,462 861,303 3.28 -3.28 0.65
2017 2013 2,677,121 1,095,950 0.092 0.118 380,434 2,296,687 6.04 -6.04 0.71
2018 2014 1,604,307 200,500 0.081 0.127 68,902 1,535,406 22.28 -22.28 0.76
2019 2015 1,230,186 5,770,650 0.004 0.311 1,384,337 154,151 0.11 0.11 0.12
2020 2016 1,839,804 1,775,820 0.016 0.264 672,195 1,167,609 1.74 -1.74 0.57
2021 2017 2,307,623 2,041,824 0.023 0.234 773,008 1,534,615 1.99 -1.99 0.59
2022 2018 NA 1,581,426 0.033 0.199 697,527 NA NA NA NA

ricker

Return Yr Brood Yr Observed Return Spawner Return R Sq P-val Forecasted Return AD APE PE AAPE
2013 2009 460,969 970,818 -0.081 0.887 237,057 223,912 0.94 -0.94 0.45
2014 2010 735,100 1,187,730 -0.076 0.919 309,982 425,119 1.37 -1.37 0.52
2015 2011 3,087,468 883,794 -0.071 0.950 243,966 2,843,502 11.66 -11.66 0.74
2016 2012 1,123,765 861,747 -0.066 0.906 291,207 832,558 2.86 -2.86 0.64
2017 2013 2,677,121 1,095,950 -0.059 0.830 402,570 2,274,551 5.65 -5.65 0.70
2018 2014 1,604,307 200,500 -0.053 0.761 90,654 1,513,654 16.70 -16.70 0.76
2019 2015 1,230,186 5,770,650 -0.032 0.534 1,078,152 152,035 0.14 -0.14 0.12
2020 2016 1,839,804 1,775,820 -0.022 0.461 775,248 1,064,556 1.37 -1.37 0.52
2021 2017 2,307,623 2,041,824 -0.020 0.452 878,903 1,428,720 1.63 -1.63 0.55
2022 2018 NA 1,581,426 -0.019 0.454 777,881 NA NA NA NA

average

Return Yr Brood Yr Observed Return Forecasted Return AD APE PE AAPE
2013 2009 460,968.7 655,280 194,311 0.30 0.30 0.40
2014 2010 735,100.4 575,386 159,715 0.28 -0.28 0.21
2015 2011 3,087,467.6 574,048 2,513,420 4.38 -4.38 0.68
2016 2012 1,123,765.3 1,058,166 65,600 0.06 -0.06 0.06
2017 2013 2,677,120.7 1,182,821 1,494,299 1.26 -1.26 0.51
2018 2014 1,604,307.4 1,616,885 12,577 0.01 0.01 0.01
2019 2015 1,230,186.2 1,845,552 615,366 0.33 0.33 0.46
2020 2016 1,839,803.8 1,944,569 104,766 0.05 0.05 0.06
2021 2017 2,307,622.8 1,695,037 612,586 0.36 -0.36 0.26
2022 2018 NA 1,931,808 NA NA NA NA

Performance summary

The average model performs best here, suggesting a forecast of ~1.9 million fish.

model range MAD MAPE MPE MAAPE
sibling 5 years 962,052 216.83% 128.91% 0.42
sibling 3 years 532,572 39.87% -34.03% 0.25
ln.sibling 5 years 1,579,981 495.21% -495.21% 0.67
ln.sibling 3 years 1,370,079 320.70% -320.70% 0.64
lnY.sibling 5 years 1,487,045 686.83% -686.83% 0.64
lnY.sibling 3 years 1,237,725 220.73% -220.73% 0.59
lnS.lnR 5 years 1,337,693 643.09% -638.64% 0.55
lnS.lnR 3 years 952,125 127.79% -120.36% 0.43
ricker 5 years 1,286,703 509.74% -509.74% 0.53
ricker 3 years 881,770 104.66% -104.66% 0.40
average 5 years 567,919 40.40% -24.59% 0.26
average 3 years 444,239 24.96% 0.86% 0.26

age.22

Below are model outputs using data from 1993

lnS.lnR

Return Yr Brood Yr Observed Return Spawner Return R Sq P-val Forecasted Return AD APE PE AAPE
2013 2008 89,650 2,180,502 0.004 0.322 62,479 27,171 0.43 -0.43 0.29
2014 2009 117,590 970,818 0.016 0.282 34,873 82,716 2.37 -2.37 0.61
2015 2010 1,800,911 1,187,730 0.012 0.290 43,770 1,757,141 40.14 -40.14 0.77
2016 2011 533,353 883,794 0.000 0.334 43,706 489,647 11.20 -11.20 0.74
2017 2012 22,311 861,747 -0.014 0.398 50,564 28,253 0.56 0.56 0.90
2018 2013 3,198 1,095,950 -0.006 0.361 56,137 52,939 0.94 0.94 1.51
2019 2014 24,204 200,500 -0.004 0.350 15,522 8,682 0.56 -0.56 0.34
2020 2015 19,770 5,770,650 0.003 0.315 138,106 118,336 0.86 0.86 1.41
2021 2016 356,462 1,775,820 -0.020 0.457 56,717 299,745 5.28 -5.28 0.70
2022 2017 NA 2,041,824 -0.012 0.402 66,426 NA NA NA NA

ricker

Return Yr Brood Yr Observed Return Spawner Return R Sq P-val Forecasted Return AD APE PE AAPE
2013 2008 89,650 2,180,502 -0.048 0.562 65,218 24,432 0.37 -0.37 0.27
2014 2009 117,590 970,818 -0.044 0.554 38,301 79,289 2.07 -2.07 0.59
2015 2010 1,800,911 1,187,730 -0.033 0.498 48,362 1,752,549 36.24 -36.24 0.77
2016 2011 533,353 883,794 -0.023 0.441 49,029 484,324 9.88 -9.88 0.74
2017 2012 22,311 861,747 -0.008 0.365 56,601 34,290 0.61 0.61 0.99
2018 2013 3,198 1,095,950 -0.010 0.377 63,007 59,810 0.95 0.95 1.52
2019 2014 24,204 200,500 -0.020 0.446 12,584 11,620 0.92 -0.92 0.45
2020 2015 19,770 5,770,650 -0.011 0.388 72,984 53,214 0.73 0.73 1.22
2021 2016 356,462 1,775,820 0.047 0.163 69,129 287,333 4.16 -4.16 0.68
2022 2017 NA 2,041,824 0.045 0.162 76,690 NA NA NA NA

average

Return Yr Brood Yr Observed Return Forecasted Return AD APE PE AAPE
2013 2008 89,650.00 155,761 66,111 0.42 0.42 0.64
2014 2009 117,589.70 172,615 55,025 0.32 0.32 0.44
2015 2010 1,800,911.37 174,188 1,626,723 9.34 -9.34 0.73
2016 2011 533,352.51 508,861 24,492 0.05 -0.05 0.05
2017 2012 22,310.67 604,573 582,263 0.96 0.96 1.53
2018 2013 3,197.53 512,763 509,565 0.99 0.99 1.56
2019 2014 24,204.11 495,472 471,268 0.95 0.95 1.52
2020 2015 19,769.62 476,795 457,026 0.96 0.96 1.53
2021 2016 356,462.27 120,567 235,895 1.96 -1.96 0.58
2022 2017 NA 85,189 NA NA NA NA

Performance summary

Nothing significant here. The average gives us a forecast of ~0.085 million.

model range MAD MAPE MPE MAAPE
sibling 5 years 198,041 89.15% 58.47% 1.27
sibling 3 years 180,481 85.44% 34.31% 1.11
ln.sibling 5 years 96,691 277.62% -177.92% 0.94
ln.sibling 3 years 145,500 419.21% -340.02% 0.88
lnY.sibling 5 years 89,607 194.22% -89.79% 0.98
lnY.sibling 3 years 120,608 271.99% -201.37% 0.80
lnS.lnR 5 years 101,591 164.06% -69.71% 0.97
lnS.lnR 3 years 142,254 223.37% -166.25% 0.82
ricker 5 years 89,253 147.28% -55.91% 0.97
ricker 3 years 117,389 193.63% -145.03% 0.78
average 5 years 451,203 116.46% 38.20% 1.35
average 3 years 388,063 128.87% -1.56% 1.21

age.13

Below are model outputs using data from 1995

sibling

Return Yr Brood Yr Observed Return Sibling Return R Sq P-val Forecasted Return AD APE PE AAPE
2013 2008 1,252,445 506,805 0.658 0.000 1,478,119 225,674 0.15 0.15 0.18
2014 2009 692,685 460,969 0.665 0.000 1,433,767 741,082 0.52 0.52 0.82
2015 2010 3,183,142 735,100 0.657 0.000 1,526,759 1,656,383 1.08 -1.08 0.48
2016 2011 2,498,861 3,087,468 0.526 0.001 2,905,360 406,499 0.14 0.14 0.16
2017 2012 1,614,661 1,123,765 0.531 0.001 1,825,170 210,509 0.12 0.12 0.13
2018 2013 1,235,290 2,677,121 0.532 0.000 2,616,545 1,381,255 0.53 0.53 0.84
2019 2014 586,591 1,604,307 0.435 0.001 1,972,258 1,385,667 0.70 0.70 1.17
2020 2015 2,693,830 1,230,186 0.376 0.002 1,730,391 963,439 0.56 -0.56 0.34
2021 2016 2,680,396 1,839,804 0.355 0.003 2,050,806 629,590 0.31 -0.31 0.23
2022 2017 NA 2,307,623 0.361 0.002 2,298,790 NA NA NA NA

ln.sibling

Return Yr Brood Yr Observed Return Sibling Return R Sq P-val Forecasted Return AD APE PE AAPE
2013 2008 1,252,445 506,805 0.318 0.026 1,738,151 485,706 0.28 0.28 0.37
2014 2009 692,685 460,969 0.304 0.024 1,676,045 983,360 0.59 0.59 0.96
2015 2010 3,183,142 735,100 0.224 0.043 1,674,980 1,508,163 0.90 -0.90 0.44
2016 2011 2,498,861 3,087,468 0.223 0.037 2,119,827 379,035 0.18 -0.18 0.15
2017 2012 1,614,661 1,123,765 0.253 0.023 1,879,377 264,716 0.14 0.14 0.16
2018 2013 1,235,290 2,677,121 0.250 0.020 2,092,991 857,702 0.41 0.41 0.61
2019 2014 586,591 1,604,307 0.207 0.029 1,872,396 1,285,806 0.69 0.69 1.14
2020 2015 2,693,830 1,230,186 0.109 0.084 1,684,818 1,009,012 0.60 -0.60 0.36
2021 2016 2,680,396 1,839,804 0.120 0.069 1,809,476 870,920 0.48 -0.48 0.31
2022 2017 NA 2,307,623 0.136 0.051 1,904,696 NA NA NA NA

lnY.sibling

Return Yr Brood Yr Observed Return Sibling Return R Sq P-val Forecasted Return AD APE PE AAPE
2013 2008 1,252,445 506,805 0.524 0.003 1,387,295 134,850 0.10 0.10 0.11
2014 2009 692,685 460,969 0.534 0.002 1,359,695 667,009 0.49 0.49 0.77
2015 2010 3,183,142 735,100 0.501 0.002 1,382,529 1,800,613 1.30 -1.30 0.51
2016 2011 2,498,861 3,087,468 0.400 0.005 2,636,079 137,218 0.05 0.05 0.05
2017 2012 1,614,661 1,123,765 0.427 0.003 1,606,785 7,876 0.00 0.00 0.00
2018 2013 1,235,290 2,677,121 0.429 0.002 2,361,348 1,126,058 0.48 0.48 0.74
2019 2014 586,591 1,604,307 0.354 0.004 1,735,830 1,149,240 0.66 0.66 1.10
2020 2015 2,693,830 1,230,186 0.257 0.013 1,513,802 1,180,028 0.78 -0.78 0.41
2021 2016 2,680,396 1,839,804 0.240 0.014 1,771,174 909,221 0.51 -0.51 0.33
2022 2017 NA 2,307,623 0.247 0.011 2,004,978 NA NA NA NA

lnS.lnR

Return Yr Brood Yr Observed Return Spawner Return R Sq P-val Forecasted Return AD APE PE AAPE
2013 2008 1,252,445 2,180,502 -0.038 0.470 1,766,736 514,291 0.29 0.29 0.39
2014 2009 692,685 970,818 -0.044 0.516 1,553,684 860,998 0.55 0.55 0.89
2015 2010 3,183,142 1,187,730 -0.025 0.431 1,501,563 1,681,579 1.12 -1.12 0.49
2016 2011 2,498,861 883,794 -0.032 0.475 1,512,413 986,449 0.65 -0.65 0.38
2017 2012 1,614,661 861,747 -0.040 0.543 1,567,143 47,518 0.03 -0.03 0.03
2018 2013 1,235,290 1,095,950 -0.036 0.531 1,614,317 379,027 0.23 0.23 0.30
2019 2014 586,591 200,500 -0.031 0.505 1,300,387 713,797 0.55 0.55 0.88
2020 2015 2,693,830 5,770,650 0.070 0.136 2,249,911 443,919 0.20 -0.20 0.16
2021 2016 2,680,396 1,775,820 0.116 0.073 1,729,812 950,584 0.55 -0.55 0.34
2022 2017 NA 2,041,824 0.125 0.059 1,837,725 NA NA NA NA

ricker

Return Yr Brood Yr Observed Return Spawner Return R Sq P-val Forecasted Return AD APE PE AAPE
2013 2008 1,252,445 2,180,502 0.638 0.001 2,332,361 1,079,917 0.46 0.46 0.71
2014 2009 692,685 970,818 0.621 0.000 1,687,798 995,113 0.59 0.59 0.96
2015 2010 3,183,142 1,187,730 0.551 0.001 1,744,122 1,439,020 0.83 -0.83 0.42
2016 2011 2,498,861 883,794 0.548 0.001 1,543,923 954,939 0.62 -0.62 0.37
2017 2012 1,614,661 861,747 0.558 0.000 1,579,259 35,402 0.02 -0.02 0.02
2018 2013 1,235,290 1,095,950 0.565 0.000 1,814,760 579,470 0.32 0.32 0.44
2019 2014 586,591 200,500 0.554 0.000 478,346 108,245 0.23 -0.23 0.18
2020 2015 2,693,830 5,770,650 0.577 0.000 1,200,214 1,493,617 1.24 -1.24 0.51
2021 2016 2,680,396 1,775,820 0.578 0.000 2,243,254 437,141 0.19 -0.19 0.16
2022 2017 NA 2,041,824 0.577 0.000 2,349,867 NA NA NA NA

average

Return Yr Brood Yr Observed Return Forecasted Return AD APE PE AAPE
2013 2008 1,252,444.7 1,941,280 688,836 0.35 0.35 0.50
2014 2009 692,685.2 1,432,526 739,841 0.52 0.52 0.82
2015 2010 3,183,142.2 1,295,460 1,887,682 1.46 -1.46 0.54
2016 2011 2,498,861.5 1,561,029 937,833 0.60 -0.60 0.36
2017 2012 1,614,660.8 1,745,225 130,564 0.07 0.07 0.08
2018 2013 1,235,289.8 1,848,359 613,069 0.33 0.33 0.46
2019 2014 586,590.7 1,844,928 1,258,337 0.68 0.68 1.13
2020 2015 2,693,830.1 1,823,709 870,121 0.48 -0.48 0.31
2021 2016 2,680,395.5 1,725,847 954,549 0.55 -0.55 0.34
2022 2017 NA 1,762,153 NA NA NA NA

Performance summary

Blending everything here gives us a forecast of ~2 million fish.

model range MAD MAPE MPE MAAPE
sibling 5 years 914,092 44.19% 9.64% 0.54
sibling 3 years 992,899 52.21% -5.37% 0.58
ln.sibling 5 years 857,631 46.35% 3.14% 0.52
ln.sibling 3 years 1,055,246 58.90% -13.12% 0.61
lnY.sibling 5 years 874,485 48.73% -3.18% 0.52
lnY.sibling 3 years 1,079,496 65.16% -21.03% 0.61
lnS.lnR 5 years 506,969 31.22% 0.13% 0.34
lnS.lnR 3 years 702,766 43.19% -6.60% 0.46
ricker 5 years 530,775 40.15% -27.37% 0.26
ricker 3 years 679,668 55.52% -55.52% 0.28
average 5 years 765,328 42.38% 1.17% 0.47
average 3 years 1,027,669 57.08% -11.61% 0.60

age.23

Below are model outputs using data from 1993

sibling

Return Yr Brood Yr Observed Return Sibling Return R Sq P-val Forecasted Return AD APE PE AAPE
2013 2007 126,186 481,363 0.580 0.001 690,973 564,787 0.82 0.82 1.35
2014 2008 70,433 89,650 0.366 0.010 155,809 85,376 0.55 0.55 0.88
2015 2009 132,401 117,590 0.373 0.007 176,762 44,361 0.25 0.25 0.32
2016 2010 766,628 1,800,911 0.376 0.005 1,786,470 1,019,842 0.57 0.57 0.93
2017 2011 48,131 533,353 0.393 0.003 360,737 312,606 0.87 0.87 1.42
2018 2012 17,181 22,311 0.342 0.005 123,009 105,828 0.86 0.86 1.41
2019 2013 9,522 3,198 0.354 0.003 108,128 98,606 0.91 0.91 1.47
2020 2014 1,068 24,204 0.365 0.002 111,023 109,955 0.99 0.99 1.56
2021 2015 697,456 19,770 0.373 0.002 102,884 594,573 5.78 -5.78 0.71
2022 2016 NA 356,462 0.268 0.007 271,012 NA NA NA NA

ln.sibling

Return Yr Brood Yr Observed Return Sibling Return R Sq P-val Forecasted Return AD APE PE AAPE
2013 2007 126,186 481,363 0.369 0.013 323,093 196,907 0.61 0.61 1.00
2014 2008 70,433 89,650 0.351 0.012 128,050 57,617 0.45 0.45 0.69
2015 2009 132,401 117,590 0.345 0.010 138,729 6,328 0.05 0.05 0.05
2016 2010 766,628 1,800,911 0.351 0.007 473,679 292,948 0.62 -0.62 0.37
2017 2011 48,131 533,353 0.422 0.002 295,089 246,958 0.84 0.84 1.38
2018 2012 17,181 22,311 0.351 0.004 62,778 45,597 0.73 0.73 1.21
2019 2013 9,522 3,198 0.358 0.003 24,608 15,086 0.61 0.61 1.01
2020 2014 1,068 24,204 0.408 0.001 56,154 55,086 0.98 0.98 1.55
2021 2015 697,456 19,770 0.329 0.003 41,406 656,050 15.84 -15.84 0.75
2022 2016 NA 356,462 0.268 0.007 189,160 NA NA NA NA

lnY.sibling

Return Yr Brood Yr Observed Return Sibling Return R Sq P-val Forecasted Return AD APE PE AAPE
2013 2007 126,186 481,363 0.256 0.037 563,870 437,684 0.78 0.78 1.29
2014 2008 70,433 89,650 0.211 0.048 71,350 918 0.01 0.01 0.01
2015 2009 132,401 117,590 0.217 0.040 80,095 52,305 0.65 -0.65 0.38
2016 2010 766,628 1,800,911 0.216 0.034 89,006,443 88,239,815 0.99 0.99 1.56
2017 2011 48,131 533,353 0.210 0.032 172,687 124,556 0.72 0.72 1.20
2018 2012 17,181 22,311 0.182 0.039 67,426 50,244 0.75 0.75 1.24
2019 2013 9,522 3,198 0.196 0.029 59,764 50,242 0.84 0.84 1.38
2020 2014 1,068 24,204 0.209 0.021 55,408 54,340 0.98 0.98 1.55
2021 2015 697,456 19,770 0.190 0.025 43,939 653,518 14.87 -14.87 0.75
2022 2016 NA 356,462 0.144 0.042 97,659 NA NA NA NA

lnS.lnR

Return Yr Brood Yr Observed Return Spawner Return R Sq P-val Forecasted Return AD APE PE AAPE
2013 2007 126,186 2,466,414 -0.077 0.801 97,390 28,796 0.30 -0.30 0.22
2014 2008 70,433 2,180,502 -0.069 0.758 98,455 28,022 0.28 0.28 0.38
2015 2009 132,401 970,818 -0.065 0.774 84,433 47,967 0.57 -0.57 0.35
2016 2010 766,628 1,187,730 -0.061 0.785 89,431 677,196 7.57 -7.57 0.72
2017 2011 48,131 883,794 -0.059 0.809 97,159 49,028 0.50 0.50 0.79
2018 2012 17,181 861,747 -0.053 0.765 92,503 75,321 0.81 0.81 1.35
2019 2013 9,522 1,095,950 -0.045 0.673 87,662 78,140 0.89 0.89 1.45
2020 2014 1,068 200,500 -0.041 0.651 52,871 51,803 0.98 0.98 1.55
2021 2015 697,456 5,770,650 0.072 0.120 238,707 458,749 1.92 -1.92 0.58
2022 2016 NA 1,775,820 0.132 0.050 104,691 NA NA NA NA

ricker

Return Yr Brood Yr Observed Return Spawner Return R Sq P-val Forecasted Return AD APE PE AAPE
2013 2007 126,186 2,466,414 0.102 0.141 126,410 224 0.00 0.00 0.00
2014 2008 70,433 2,180,502 0.111 0.122 127,066 56,633 0.45 0.45 0.68
2015 2009 132,401 970,818 0.120 0.103 94,492 37,908 0.40 -0.40 0.28
2016 2010 766,628 1,187,730 0.131 0.085 107,143 659,485 6.16 -6.16 0.71
2017 2011 48,131 883,794 0.136 0.073 104,517 56,386 0.54 0.54 0.86
2018 2012 17,181 861,747 0.129 0.073 97,717 80,535 0.82 0.82 1.36
2019 2013 9,522 1,095,950 0.101 0.094 100,263 90,741 0.91 0.91 1.47
2020 2014 1,068 200,500 0.072 0.127 23,358 22,290 0.95 0.95 1.52
2021 2015 697,456 5,770,650 0.014 0.268 93,672 603,784 6.45 -6.45 0.71
2022 2016 NA 1,775,820 -0.024 0.491 107,638 NA NA NA NA

average

Return Yr Brood Yr Observed Return Forecasted Return AD APE PE AAPE
2013 2007 126,186.130 90,890 35,296 0.39 -0.39 0.27
2014 2008 70,432.886 108,446 38,013 0.35 0.35 0.49
2015 2009 132,400.541 118,245 14,155 0.12 -0.12 0.11
2016 2010 766,627.812 115,057 651,570 5.66 -5.66 0.70
2017 2011 48,130.986 246,763 198,632 0.80 0.80 1.33
2018 2012 17,181.106 228,756 211,575 0.92 0.92 1.49
2019 2013 9,521.967 206,955 197,433 0.95 0.95 1.52
2020 2014 1,067.886 194,772 193,705 0.99 0.99 1.57
2021 2015 697,456.382 168,506 528,950 3.14 -3.14 0.65
2022 2016 NA 154,672 NA NA NA NA

Performance summary

Blending everything but the ricker gives us ~0.16 million.

model range MAD MAPE MPE MAAPE
sibling 5 years 244,313 188.17% -43.00% 1.31
sibling 3 years 267,711 256.05% -129.23% 1.25
ln.sibling 5 years 203,756 380.03% -253.74% 1.18
ln.sibling 3 years 242,074 581.28% -475.01% 1.10
lnY.sibling 5 years 186,580 363.22% -231.71% 1.23
lnY.sibling 3 years 252,700 556.49% -435.07% 1.23
lnS.lnR 5 years 142,608 102.24% 25.37% 1.14
lnS.lnR 3 years 196,231 126.43% -1.69% 1.19
ricker 5 years 170,747 193.37% -64.46% 1.19
ricker 3 years 238,938 276.83% -152.88% 1.23
average 5 years 266,059 136.35% 10.79% 1.31
average 3 years 306,696 169.59% -39.69% 1.25

Forecast summary

A preliminary look at the models for this river leads us to an initial forecast of ~4.2 million broken down as follows:

Table 2. Preliminary Forecast.

age forecast
1.2 1.900
2.2 0.085
1.3 2.000
2.3 0.160