# Confidence Interval Estimation for Precipitation Quantiles Based on Principle of Maximum Entropy

^{*}

## Abstract

**:**

## 1. Introduction

## 2. Confidence Interval Estimation of the Quantile

#### 2.1. Estimation of Quantile

#### 2.2. Calculation of Confidence Interval

#### 2.2.1. Method of Moments (MOM)

#### 2.2.2. Maximum Likelihood (ML) Method

**I**[28]:

#### 2.2.3. Principle of Maximum Entropy (POME) Method

## 3. Asymptotic Variances of Quantile Estimators for Different Distributions

#### 3.1. Gamma Distribution

#### 3.1.1. Estimation of Asymptotic Variances by MOM and ML

#### 3.1.2. Estimation of Asymptotic Variances by POME

#### 3.2. Pearson Type 3 (P3) Distribution

#### 3.2.1. Estimation of Asymptotic Variances by MOM and ML

#### 3.2.2. Estimation of Asymptotic Variances by POME

#### 3.3. Extreme Value Type 1 (EV1) Distribution

#### 3.3.1. Estimation of asymptotic variance by MOM and ML

#### 3.3.2. Estimation of Asymptotic Variances by POME

## 4. Simulation Experiments

## 5. Application

## 6. Conclusions

- (1)
- The calculation formulas of the asymptotic variances and confidence intervals of quantiles for three distributions based on POME are given. The results of simulation experiments and the case study show that the POME method can provide an effective way for reducing the uncertainty of quantile estimators.
- (2)
- Results of the simulation experiments demonstrate that the POME method yields the smallest standard errors and the narrowest confidence intervals of quantile estimators compared with the results of MOM and ML. This may benefit from fewer sampling errors and approximation in derivation. Thus, the POME can give more accurate estimates. Furthermore, the standard errors and confidence interval widths of the quantiles increased with the return period T and decreased with the sample size.
- (3)
- Results of the case study indicate that when using different criteria for distribution selection, the results are coincidental, and the POME is the optimal method for parameter estimation. Furthermore, the POME can give more reliable precipitation quantiles since the standard errors and 95% confidence interval widths of precipitation quantiles obtained by POME are smaller than those obtained by the MOM and the ML methods.

## Author Contributions

## Funding

## Acknowledgments

## Conflicts of Interest

## Appendix A

#### Estimation of $\mathrm{var}\left(\overline{W}\right)$

#### Estimation of $\mathrm{cov}\left(\overline{X},\overline{W}\right)$

## References

- Hamed, K.; Rao, A.R. Flood Frequency Analysis; Taylor & Francis Inc.: Bosa Roca, FL, USA, 1999. [Google Scholar]
- Obeysekera, J.; Salas, J.D. Quantifying the uncertainty of design floods under nonstationary conditions. J. Hydrol. Eng.
**2014**, 19, 1438–1446. [Google Scholar] [CrossRef] - Chowdhury, J.U.; Stedinger, J.R. Confidence interval for design floods with estimated skew coefficient. J. Hydraul. Eng.
**1991**, 117, 811–831. [Google Scholar] [CrossRef] - Lu, L.H.; Stedinger, J.R. Sampling variance of normalized GEV/PWM quantile estimators and a regional homogeneity test. J. Hydrol.
**1992**, 138, 223–245. [Google Scholar] [CrossRef] - Hoshi, K.; Burges, S.J. Sampling properties of parameter estimates for the log Pearson type 3 distribution, using moments in real space. J. Hydrol.
**1981**, 53, 305–316. [Google Scholar] [CrossRef] - Condie, R. The Log Pearson Type 3 Distribution: The T-Year Event and Its Asymptotic Standard Error by Maximum Likelihood Theory. Water Resour. Res.
**1977**, 13, 987–991. [Google Scholar] [CrossRef] - Phien, H.N. Variance of the T-year event in the log Pearson type-3 distribution. J. Hydrol.
**1985**, 77, 141–158. [Google Scholar] [CrossRef] - Heo, J.H.; Salas, J.D. Estimation of quantiles and confidence intervals for the log-Gumbel distribution. Stoch. Hydrol. Hydraul.
**1996**, 10, 187–207. [Google Scholar] [CrossRef] - Heo, J.H.; Salas, D.J.; Kim, K.D. Estimation of confidence intervals of quantiles for the Weibull distribution. Stoch. Environ. Res. Risk Assess.
**2001**, 15, 284–309. [Google Scholar] [CrossRef] - Shin, H.; Kim, T.; Kim, S.; Heo, J.H. Estimation of asymptotic variances of quantiles for the generalized logistic distribution. Stoch. Environ. Res. Risk Assess.
**2010**, 24, 183–197. [Google Scholar] [CrossRef] - Shannon, E.C. A mathematical theory of communication. Bell Labs Tech. J.
**1948**, 4, 379–423. [Google Scholar] [CrossRef] - Jaynes, E.T. On the rationale of maximum-entropy methods. Proc. IEEE
**1982**, 70, 939–952. [Google Scholar] [CrossRef] - Sonuga, J.O. Principle of maximum entropy in hydrologic frequency analysis. J. Hydrol.
**1972**, 17, 177–191. [Google Scholar] [CrossRef] - Singh, V.P.; Rajagopal, A.K.; Singh, K. Derivation of some frequency distributions using the principle of maximum entropy (POME). Adv. Water Resour.
**1986**, 9, 91–106. [Google Scholar] [CrossRef] - Lu, C.; Singh, V.P. Entropy-based derivation of generalized distributions for hydrometeorological frequency analysis. J. Hydrol.
**2018**, 557, 699–712. [Google Scholar] - Chen, L.; Singh, V.; Xiong, F. An Entropy-Based Generalized Gamma Distribution for Flood Frequency Analysis. Entropy
**2017**, 19, 239. [Google Scholar] [CrossRef] - Singh, V. Entropy-Based Parameter Estimation in Hydrology; Kluwer Academic Publishers: Boston, MA, USA, 1998. [Google Scholar]
- Singh, V.P.; Deng, Z.Q. Entropy-Based Parameter Estimation for Kappa Distribution. J. Hydrol. Eng.
**2003**, 8, 81–92. [Google Scholar] [CrossRef] - Song, S.; Song, X.; Kang, Y. Entropy-Based Parameter Estimation for the Four-Parameter Exponential Gamma Distribution. Entropy
**2017**, 19, 189. [Google Scholar] [CrossRef] - Fiorentino, M.; Arora, K.; Singh, V.P. The two-component extreme value distribution for flood frequency analysis: Derivation of a new estimation method. Stoch. Hydrol. Hydraul.
**1987**, 1, 199–208. [Google Scholar] [CrossRef] - Hao, Z.; Singh, V.P. Modeling multisite streamflow dependence with maximum entropy copula. Water Resour. Res.
**2013**, 49, 7139–7143. [Google Scholar] [CrossRef][Green Version] - Kong, X.M.; Huang, G.H.; Fan, Y.R.; Li, Y.P. Maximum entropy-Gumbel-Hougaard copula method for simulation of monthly streamflow in Xiangxi river, China. Stoch. Env. Res. Risk Assess.
**2015**, 29, 833–846. [Google Scholar] [CrossRef] - Aghakouchak, A. Entropy—Copula in Hydrology and Climatology. J. Hydrometeorol.
**2014**, 15, 2176–2189. [Google Scholar] [CrossRef] - Phien, H.N.; Nguyen, V.T.V. Variances and covariances of the maximum entropy estimators for the Pearson type-3 distribution. Can. J. Civ. Eng.
**1990**, 17, 590–596. [Google Scholar] [CrossRef] - Phien, H.N. Sampling properties of the maximum entropy estimators for the extreme-value type-1 distribution. J. Hydrol.
**1986**, 86, 391–398. [Google Scholar] [CrossRef] - Chow, V.T. A General Formula for Hydrologic Frequency Analysis. Eos Trans. Am. Geophys. Union
**1951**, 32, 231–237. [Google Scholar] [CrossRef] - Kite, G.W. Confidence Limits for Design Events. Water Resour. Res.
**1975**, 11, 48–53. [Google Scholar] [CrossRef] - Chowdhary, H.; Singh, V.P. Reducing uncertainty in estimates of frequency distribution parameters using composite likelihood approach and copula-based bivariate distributions. Water Resour. Res.
**2010**, 46. [Google Scholar] [CrossRef][Green Version] - Bobee, B. Sample error of T-year events commuted by fitting a Pearson type 3 distribution. Water Resour. Res.
**1973**, 9, 1264–1270. [Google Scholar] [CrossRef] - Strupczewski, W.G.; Kochanek, K.; Singh, V.P.; Węglarczyk, S. Are parsimonious FF models more reliable than true ones? II. Comparative assessment of the performance of simple models versus the parent distributions. Acta Geophys. Pol.
**2005**, 53, 437–457. [Google Scholar] - Houghton, J.C. Birth of a Parent: The Wakeby Distribution for Modeling Flood Flows. Water Resour. Res.
**1978**, 14, 1105–1109. [Google Scholar] [CrossRef]

Case | ξ | α | β | γ | δ | Cv | Cs |
---|---|---|---|---|---|---|---|

I | 0 | 16 | 16 | 0.4 | 0.04 | 0.38 | 1.10 |

I | 15.4 | 308.8 | 10.25 | 38.5 | −0.30 | 0.36 | 0.48 |

III | 273.69 | 521.10 | 1.25 | 4.77 | −0.21 | 0.24 | 0.64 |

**Table 2.**Median of estimated quantiles (${\widehat{x}}_{T}$), standard error (St), and 95% confidence interval (CI) width from generated data; MOM = methods of moments, ML = maximum likelihood, POME = principle of maximum entropy.

Case | Sample Size n | Return Period T | MOM | ML | POME | ||||||
---|---|---|---|---|---|---|---|---|---|---|---|

${\widehat{\mathit{x}}}_{\mathit{T}}$ | St | CI Width | ${\widehat{\mathit{x}}}_{\mathit{T}}$ | St | CI Width | ${\widehat{\mathit{x}}}_{\mathit{T}}$ | St | CI Width | |||

I | 20 | 10 | 1.97 | 0.22 | 0.85 | 2.08 | 0.22 | 0.85 | 2.02 | 0.20 | 0.80 |

100 | 2.83 | 0.41 | 1.60 | 3.06 | 0.38 | 1.49 | 2.97 | 0.35 | 1.37 | ||

200 | 3.09 | 0.47 | 1.83 | 3.36 | 0.43 | 1.69 | 3.26 | 0.40 | 1.55 | ||

50 | 10 | 2.00 | 0.14 | 0.56 | 2.13 | 0.15 | 0.58 | 2.06 | 0.13 | 0.52 | |

100 | 2.89 | 0.27 | 1.05 | 3.20 | 0.26 | 1.01 | 3.05 | 0.23 | 0.89 | ||

200 | 3.14 | 0.31 | 1.20 | 3.52 | 0.29 | 1.14 | 3.35 | 0.26 | 1.00 | ||

100 | 10 | 2.01 | 0.10 | 0.40 | 2.16 | 0.11 | 0.42 | 2.08 | 0.10 | 0.37 | |

100 | 2.93 | 0.19 | 0.76 | 3.24 | 0.19 | 0.73 | 3.09 | 0.16 | 0.64 | ||

200 | 3.20 | 0.22 | 0.87 | 3.56 | 0.21 | 0.82 | 3.39 | 0.18 | 0.72 | ||

1000 | 10 | 2.02 | 0.03 | 0.13 | 2.17 | 0.03 | 0.13 | 2.09 | 0.03 | 0.12 | |

100 | 2.95 | 0.06 | 0.24 | 3.27 | 0.06 | 0.23 | 3.12 | 0.05 | 0.20 | ||

200 | 3.23 | 0.07 | 0.28 | 3.60 | 0.07 | 0.27 | 3.42 | 0.06 | 0.23 | ||

II | 20 | 10 | 106.9 | 12.1 | 47.3 | 111.1 | 11.9 | 46.5 | 109.7 | 11.4 | 44.7 |

100 | 155.3 | 22.8 | 89.3 | 164.7 | 20.7 | 81.3 | 161.3 | 19.7 | 77.2 | ||

200 | 169.8 | 26.0 | 102.1 | 180.4 | 23.4 | 91.9 | 176.8 | 22.2 | 87.1 | ||

50 | 10 | 106.9 | 12.1 | 47.3 | 112.3 | 7.7 | 30.2 | 110.3 | 7.4 | 28.9 | |

100 | 155.3 | 22.8 | 89.3 | 167.8 | 13.5 | 52.8 | 163.7 | 12.8 | 50.0 | ||

200 | 169.8 | 26.0 | 102.1 | 184.4 | 15.2 | 59.7 | 179.4 | 14.4 | 56.5 | ||

100 | 10 | 107.1 | 5.5 | 21.5 | 113.0 | 5.5 | 21.6 | 110.8 | 5.3 | 20.6 | |

100 | 156.3 | 10.3 | 40.5 | 169.0 | 9.6 | 37.7 | 164.8 | 9.1 | 35.7 | ||

200 | 170.9 | 11.8 | 46.4 | 185.6 | 10.9 | 42.6 | 180.8 | 10.3 | 40.3 | ||

1000 | 10 | 107.5 | 1.7 | 6.8 | 113.5 | 1.8 | 6.9 | 111.2 | 1.7 | 6.6 | |

100 | 156.7 | 3.3 | 12.9 | 169.9 | 3.1 | 12.0 | 165.5 | 2.9 | 11.3 | ||

200 | 171.3 | 3.8 | 14.7 | 186.6 | 3.5 | 13.6 | 181.6 | 3.3 | 12.8 | ||

III | 20 | 10 | 676.4 | 58.1 | 227.6 | 685.7 | 57.2 | 224.1 | 698.5 | 57.2 | 224.4 |

100 | 905.4 | 109.5 | 429.3 | 944.9 | 99.9 | 391.8 | 958.1 | 99.0 | 388.0 | ||

200 | 974.0 | 125.2 | 490.9 | 1023.0 | 113.0 | 442.9 | 1036.1 | 111.6 | 437.6 | ||

50 | 10 | 675.0 | 36.7 | 143.9 | 700.5 | 37.8 | 148.4 | 700.2 | 36.8 | 144.3 | |

100 | 905.5 | 69.2 | 271.4 | 971.7 | 66.2 | 259.4 | 965.6 | 63.5 | 248.8 | ||

200 | 974.3 | 79.2 | 310.3 | 1053.7 | 74.8 | 293.3 | 1044.3 | 71.7 | 280.9 | ||

100 | 10 | 1065.5 | 92.4 | 362.0 | 1160.3 | 86.3 | 338.2 | 1148.5 | 82.5 | 323.5 | |

100 | 674.8 | 26.0 | 102.0 | 707.7 | 65.5 | 106.9 | 700.2 | 26.2 | 102.7 | ||

200 | 906.9 | 49.1 | 192.4 | 985.3 | 41.5 | 186.9 | 968.7 | 45.2 | 177.4 | ||

1000 | 10 | 975.7 | 56.1 | 220.0 | 1067.2 | 47.7 | 211.3 | 1048.4 | 51.1 | 200.2 | |

100 | 674.7 | 8.2 | 32.2 | 713.6 | 8.7 | 34.2 | 701.4 | 8.3 | 32.6 | ||

200 | 906.9 | 15.5 | 60.8 | 994.1 | 15.3 | 59.8 | 971.3 | 14.4 | 56.3 |

Station Name | Record Length (Year) | Mean (mm) | Standard Deviation | Coefficient of Variation | Skewness | First-Order Serial Correlation Coefficient |
---|---|---|---|---|---|---|

Changwu | 51 | 580.6 | 131.8177 | 0.2270 | 0.5070 | 3.2153 |

Lintong | 50 | 579.5 | 129.2014 | 0.2230 | 0.6299 | 3.7670 |

Meixian | 50 | 578.0 | 129.7214 | 0.2245 | 0.5828 | 3.4614 |

Tongguan | 52 | 605.5 | 143.4648 | 0.2369 | 0.5771 | 3.6438 |

**Table 4.**Parameter values of each distribution estimated by the three methods; P3 = Pearson type 3, EV1 = extreme value type 1.

Station Name | Method | Gamma | P3 | EV1 | ||||
---|---|---|---|---|---|---|---|---|

α | β | α | β | γ | α | u | ||

Changwu | MOM | 29.9282 | 19.3993 | 15.5623 | 33.4147 | 60.5782 | 102.7783 | 521.2619 |

ML | 29.0510 | 19.9851 | 16.0097 | 32.5878 | 58.8660 | 114.9989 | 518.3959 | |

POME | 29.3458 | 19.7893 | 11.1017 | 39.5620 | 141.3794 | 111.0142 | 516.5091 | |

Lintong | MOM | 28.8063 | 20.1169 | 10.0804 | 40.6938 | 169.2817 | 100.7383 | 521.3449 |

ML | 27.6873 | 20.9299 | 17.1488 | 30.6908 | 53.1797 | 113.0664 | 519.0634 | |

POME | 28.6849 | 20.2192 | 10.3139 | 40.2305 | 164.5577 | 108.5086 | 516.8608 | |

Meixian | MOM | 29.1161 | 19.8498 | 11.7762 | 37.8015 | 132.7913 | 101.1438 | 519.5689 |

ML | 28.0875 | 20.5768 | 17.7271 | 30.3383 | 40.1383 | 113.7362 | 517.1681 | |

POME | 28.7575 | 20.1089 | 10.8002 | 39.4726 | 151.6378 | 109.1477 | 514.9499 | |

Tongguan | MOM | 33.9927 | 17.8122 | 12.0090 | 41.3991 | 108.3228 | 111.8595 | 540.9202 |

ML | 32.9535 | 18.3740 | 15.0343 | 36.5741 | 55.6191 | 125.0433 | 537.9242 | |

POME | 33.7243 | 17.9653 | 10.3756 | 44.5388 | 143.3688 | 120.6552 | 535.8444 |

**Table 5.**Ordinary least square (OLS), Akaike information criterion (AIC) and quasi-optimal deterministic coefficient test (QD) values of three distributions calculated by MOM, ML, and POME.

Station Name | Method | Gamma | P3 | EV1 | ||||||
---|---|---|---|---|---|---|---|---|---|---|

OLS | AIC | QD | OLS | AIC | QD | OLS | AIC | QD | ||

Changwu | MOM | 16.6696 | 292.9861 | 0.9837 | 16.3741 | 291.1613 | 0.9843 | 19.3927 | 308.4197 | 0.9779 |

ML | 17.5392 | 298.1729 | 0.9819 | 17.0194 | 295.1041 | 0.983 | 16.8854 | 294.2978 | 0.9833 | |

POME | 17.2128 | 296.2565 | 0.9826 | 16.0997 | 289.4375 | 0.9848 | 16.6415 | 292.8135 | 0.9837 | |

Lintong | MOM | 19.6949 | 304.0361 | 0.9763 | 18.516 | 297.8638 | 0.979 | 18.8765 | 299.792 | 0.9782 |

ML | 20.6502 | 308.7725 | 0.9739 | 20.1246 | 306.1944 | 0.9752 | 16.3975 | 285.7129 | 0.9836 | |

POME | 19.7466 | 304.2982 | 0.9762 | 18.5442 | 298.0155 | 0.979 | 16.0775 | 283.742 | 0.9842 | |

Meixian | MOM | 17.9613 | 294.8219 | 0.9804 | 16.9804 | 289.2061 | 0.9825 | 17.8405 | 294.1469 | 0.9807 |

ML | 18.9359 | 300.106 | 0.9783 | 18.5121 | 297.8426 | 0.9792 | 15.0757 | 277.3083 | 0.9862 | |

POME | 18.2446 | 296.3869 | 0.9798 | 16.8553 | 288.4664 | 0.9828 | 14.6984 | 274.7736 | 0.9869 | |

Tongguan | MOM | 20.419 | 319.7126 | 0.9793 | 20.0141 | 317.6294 | 0.9802 | 21.9677 | 327.3156 | 0.9761 |

ML | 21.2629 | 323.9242 | 0.9776 | 20.8105 | 321.6875 | 0.9785 | 19.5845 | 315.373 | 0.981 | |

POME | 20.5825 | 320.542 | 0.979 | 19.9542 | 317.3178 | 0.9803 | 19.2731 | 313.7059 | 0.9816 |

**Table 6.**Quantile estimators, standard error, and 95% confidence interval widths based on MOM, ML, and POME for the annual precipitation (mm).

Station Name | Best fitted Distribution | Return Period (Year) | MOM | ML | POME | ||||||
---|---|---|---|---|---|---|---|---|---|---|---|

Quantile | Standard Error | Confidence Interval Width | Quantile | Standard Error | Confidence Interval width | Quantile | Standard Error | Confidence Interval Width | |||

Changwu | P3 | 10 | 755.0 | 31.2 | 122.1 | 753.1 | 30.5 | 119.6 | 755.7 | 31.7 | 124.3 |

20 | 814.7 | 40.7 | 159.7 | 811.9 | 38.9 | 152.3 | 817.5 | 38.4 | 150.7 | ||

50 | 885.6 | 56.1 | 220.0 | 881.9 | 51.7 | 202.5 | 891.6 | 47.3 | 185.6 | ||

100 | 935.3 | 69.1 | 270.7 | 930.8 | 62.2 | 243.8 | 943.8 | 54.0 | 211.6 | ||

200 | 982.3 | 82.8 | 324.5 | 977.1 | 73.3 | 287.2 | 993.5 | 60.6 | 237.5 | ||

500 | 1041.4 | 101.8 | 398.9 | 1035.3 | 88.5 | 347.0 | 1056.3 | 69.2 | 271.2 | ||

Lintong | EV1 | 10 | 704.4 | 31.8 | 124.7 | 724.5 | 32.2 | 126.1 | 714.0 | 30.6 | 120.0 |

20 | 748.0 | 37.5 | 146.9 | 773.5 | 37.0 | 144.9 | 761.0 | 35.0 | 137.1 | ||

50 | 820.6 | 47.4 | 185.7 | 854.9 | 45.2 | 177.3 | 839.2 | 42.5 | 166.7 | ||

100 | 984.8 | 70.7 | 277.1 | 1039.2 | 64.6 | 253.3 | 1016.0 | 60.3 | 236.2 | ||

200 | 1054.8 | 80.8 | 316.9 | 1117.8 | 73.1 | 286.4 | 1091.5 | 68.0 | 266.5 | ||

500 | 1147.3 | 94.3 | 369.7 | 1221.6 | 84.3 | 330.3 | 1191.1 | 78.2 | 306.6 | ||

Meixian | EV1 | 10 | 703.3 | 31.9 | 125.2 | 723.8 | 32.4 | 126.9 | 713.3 | 30.8 | 120.7 |

20 | 747.2 | 37.6 | 147.5 | 773.1 | 37.2 | 145.8 | 760.6 | 35.2 | 137.9 | ||

50 | 820.0 | 47.6 | 186.4 | 855.0 | 45.5 | 178.3 | 839.1 | 42.8 | 167.6 | ||

100 | 984.8 | 71.0 | 278.3 | 1040.4 | 65.0 | 254.8 | 1017.0 | 60.6 | 237.6 | ||

200 | 1055.2 | 81.2 | 318.2 | 1119.5 | 73.5 | 288.1 | 1093.0 | 68.4 | 268.0 | ||

500 | 1148.0 | 94.7 | 371.2 | 1223.9 | 84.8 | 332.3 | 1193.2 | 78.7 | 308.4 | ||

Tongguan | EV1 | 10 | 744.2 | 34.6 | 135.7 | 765.1 | 34.9 | 136.8 | 755.1 | 33.5 | 131.5 |

20 | 792.6 | 40.8 | 160.0 | 819.3 | 40.1 | 157.1 | 807.4 | 38.4 | 150.4 | ||

50 | 873.2 | 51.6 | 202.2 | 909.3 | 49.0 | 192.2 | 894.2 | 46.7 | 183.1 | ||

100 | 1055.5 | 77.0 | 301.8 | 1113.1 | 70.1 | 274.7 | 1090.9 | 66.3 | 259.9 | ||

200 | 1133.3 | 88.0 | 345.1 | 1200.1 | 79.2 | 310.6 | 1174.8 | 74.8 | 293.3 | ||

500 | 1236.0 | 102.7 | 402.5 | 1314.9 | 91.4 | 358.2 | 1285.5 | 86.1 | 337.6 |

**Table 7.**Change in the uncertainty in quantile estimators based on POME compared with the MOM and ML methods (%).

Station Name | Return Period(Year) | POME to MOM | POME to ML | ||||
---|---|---|---|---|---|---|---|

Quantile | Standard Error | Confidence Interval Width | Quantile | Standard Error | Confidence Interval Width | ||

Changwu | 10 | 0.09 | 1.74 | 1.74 | 0.34 | 3.78 | 3.78 |

20 | 0.34 | −5.63 | −5.63 | 0.68 | −1.02 | −1.02 | |

50 | 0.67 | −15.66 | −15.66 | 1.10 | −7.68 | −7.68 | |

100 | 0.91 | −21.83 | −21.83 | 1.39 | −11.88 | −11.88 | |

200 | 1.14 | −26.82 | −26.82 | 1.67 | −15.33 | −15.33 | |

500 | 1.43 | −32.02 | −32.02 | 2.02 | −18.99 | −18.99 | |

Lintong | 10 | 1.37 | −3.74 | −3.74 | −1.49 | −4.93 | −4.93 |

20 | 1.74 | −6.67 | −6.67 | −1.67 | −5.29 | −5.29 | |

50 | 2.27 | −10.23 | −10.23 | −1.92 | −5.70 | −5.70 | |

100 | 3.17 | −14.76 | −14.76 | −2.35 | −6.18 | −6.18 | |

200 | 3.48 | −15.92 | −15.92 | −2.50 | −6.29 | −6.29 | |

500 | 3.82 | −17.05 | −17.05 | −2.66 | −6.41 | −6.41 | |

Meixian | 10 | 1.41 | −3.58 | −3.58 | −1.50 | −4.95 | −4.95 |

20 | 1.79 | −6.51 | −6.51 | −1.68 | −5.32 | −5.32 | |

50 | 2.34 | −10.08 | −10.08 | −1.93 | −5.73 | −5.73 | |

100 | 3.27 | −14.62 | −14.62 | −2.37 | −6.21 | −6.21 | |

200 | 3.58 | −15.78 | −15.78 | −2.51 | −6.33 | −6.33 | |

500 | 3.93 | −16.92 | −16.92 | −2.68 | −6.44 | −6.44 | |

Tongguan | 10 | 1.47 | −3.14 | −3.14 | −1.35 | −3.92 | −3.92 |

20 | 1.86 | −5.98 | −5.98 | −1.51 | −4.20 | −4.20 | |

50 | 2.41 | −9.45 | −9.45 | −1.73 | −4.53 | −4.53 | |

100 | 3.35 | −13.88 | −13.88 | −2.11 | −4.92 | −4.92 | |

200 | 3.66 | −15.00 | −15.00 | −2.23 | −5.02 | −5.02 | |

500 | 4.01 | −16.12 | −16.12 | −2.37 | −5.11 | −5.11 |

© 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

## Share and Cite

**MDPI and ACS Style**

Wei, T.; Song, S. Confidence Interval Estimation for Precipitation Quantiles Based on Principle of Maximum Entropy. *Entropy* **2019**, *21*, 315.
https://doi.org/10.3390/e21030315

**AMA Style**

Wei T, Song S. Confidence Interval Estimation for Precipitation Quantiles Based on Principle of Maximum Entropy. *Entropy*. 2019; 21(3):315.
https://doi.org/10.3390/e21030315

**Chicago/Turabian Style**

Wei, Ting, and Songbai Song. 2019. "Confidence Interval Estimation for Precipitation Quantiles Based on Principle of Maximum Entropy" *Entropy* 21, no. 3: 315.
https://doi.org/10.3390/e21030315