Selection of appropriate time scale with Boruta algorithm for regional drought monitoring using multi-scaler drought index

Zulfiqar Ali; Ijaz Hussain; Muhammad Faisal; Elsayed Elsherbini Elashkar; Showkat Gani; Muhammad Ahmed Shehzad

doi:10.1080/16000870.2019.1604057

Selection of appropriate time scale with Boruta algorithm for regional drought monitoring using multi-scaler drought index

Tellus

Volume 71 (2019): Issue 1

By: Zulfiqar Ali, Ijaz Hussain, Muhammad Faisal, Elsayed Elsherbini Elashkar, Showkat Gani and Muhammad Ahmed Shehzad

Open Access

|Jan 2019

Figures & Tables

Table 1.

Drought classification criteria of SPTI.

Range of SPTI valuesDrought classes

\geq 2

Extremely wet (EW)1.5 to 1.99Severely wet (SW)

1 to 1.49

Moderate wet (MW)

0.99 to - 0.99

Near normal (NN)

- 1 to - 1.49

Moderate drought (MD)

- 1.5 to - 1.99

Severe drought (SD)

\leq - 2

Extreme drought (ED)

Table 2.

CDF of selected distribution for various time scale of SPTI at Chirat station.

Time scaleDistributionCDFAICDAI-13P Weibul distribution

F (x) = {_{0 x < 0}^{1 - e^{- {(x / λ)}^{k}} x \geq 0}

508.519DAI-3Generalized extreme value distribution

F (x) = e^{- t (x)}, where t (x) = {_{e^{- (x - u) / σ} iff ξ = 0}^{{(1 + ξ (\frac{x - u}{σ}))}^{\frac{- 1}{ξ}} iff ξ \neq 0}

542.364DAI-6Generalized trapezoida distributionComplex: see van Dorp and Kotz (2003)691.105DAI-9Gumbel distribution

F (x) = e^{- e^{- (x - u) / β}}

827.091DAI-12Generalized trapezoidal distributionComplex: see van Dorp and Kotz (2003)997.765DAI-24Johnson’s SU-distribution

F (x) = Φ (γ + δ \sinh^{- 1} (\frac{x - ζ}{λ}))

777.716DAI-48Laplace distribution

F (x) = {_{1 - \frac{1}{2} exp (\frac{x - u}{σ}) if x < u}^{\frac{1}{2} exp (\frac{x - u}{σ}) if x \geq u}

1058.08

Temporal plot of SPTI value at various Time-Scales.

Correlation among various SPTI Time-Scales.

Boruta using normal permutation adapter.

Table 3.

Attribute statistics using normal permutation adapter.

Time scalemeanImpmedianImpminImpmaxImpnormHitsDecisionSPTI-10.9430560.93750.9319440.9708331ConfirmedSPTI-30.5270830.5284720.5145830.5319441ConfirmedSPTI-60.3965280.3965280.3909720.4076391ConfirmedSPTI-90.7173610.7180560.7034720.7270831ConfirmedSPTI-120.3361110.3368060.3250.3423611ConfirmedSPTI-240.2888890.2895830.2840280.2923611ConfirmedSPTI-480.60.60.5951390.6076391Confirmed

Table 4.

Attribute statistics using raw permutation adapter.

Time scalemeanImpmedianImpminImpmaxImpnormHitsDecisionSPTI-15462.075455.705410.075545.371ConfirmedSPTI-33174.623174.603157.793189.491ConfirmedSPTI-62396.072395.722392.162400.241ConfirmedSPTI-94081.734083.404068.504096.211ConfirmedSPTI-122311.932312.252304.672322.541ConfirmedSPTI-242295.082293.002282.342313.871ConfirmedSPTI-483868.353865.603853.093890.851Confirmed

Table 5.

Attribute statistics using Gini importance measure.

Time scalemeanImpmedianImpminImpmaxImpnormHitsDecisionSPTI-199.4099.5494.67103.461ConfirmedSPTI-3106.19105.55101.87112.211ConfirmedSPTI-688.8688.0285.7197.411ConfirmedSPTI-9147.96148.41142.56151.701ConfirmedSPTI-1275.7175.5469.8173.641ConfirmedSPTI-2468.4368.3065.9471.521ConfirmedSPTI-48137.37136.75134.54142.151Confirmed

References

Authors

Metrics

Articles in this issue

DOI: https://doi.org/10.1080/16000870.2019.1604057 | Journal eISSN: 3035-9554

Journal RSS Feed

Language: English

Page range: 1604057 - 1604057

Published on: Jan 1, 2019

Published by: Stockholm University Press

In partnership with: Paradigm Publishing Services

Keywords:

multi-scalar drought index,

drought monitoring,

random forest

© 2019 Zulfiqar Ali, Ijaz Hussain, Muhammad Faisal, Elsayed Elsherbini Elashkar, Showkat Gani, Muhammad Ahmed Shehzad, published by Stockholm University Press
This work is licensed under the Creative Commons Attribution 4.0 License.

Volume 71 (2019): Issue 1

Selection of appropriate time scale with Boruta algorithm for regional drought monitoring using multi-scaler drought index

Figures & Tables

Table 1.

Fig. 1.

Table 2.

Fig. 3.

Fig. 4.

Fig. 5.

Fig. 6.

Fig. 7.

Table 3.

Table 4.

Table 5.

Paradigm

My account