Semarang Tidal Floods & Informal Settlements

background-image: url("img/semarang_cover.jpeg")
background-size: cover
background-position: center

# Semarang Tidal Floods & Informal Settlements
## CASA0023<br/>Group Project
### Hilman et al.

### 21/03/2023

---
class: inverse, center, middle

# Problem: Context & Background

---
class: center, middle

## Semarang suffers from tidal floods every year

.pull-left[
<img src="img/s3_img1.png" width="120%" style="display: block; margin: auto auto auto 0;" />
]
.pull-right[
<img src="img/s3_img2.png" width="120%" style="display: block; margin: auto auto auto 0;" />
]

Image credit: [Merah Putih](https://merahputih.com/post/read/banjir-rob-genangi-sejumlah-wilayah-pesisir-di-semarang), [Merdekah](https://www.merdeka.com/jateng/5-fakta-banjir-rob-semarang-ketinggian-air-capai-15-meter.html)

---
# Location

- Semarang is capital city of Central Java, Indonesia

- **North of Semarang is located next to Java Sea** which become the main port and access of food supply  for Central Java.

Source: <a name=cite-dameriaRelationshipResidentsSense2022></a>[Dameria, Akbar, Indradjati, and Tjokropandojo (2022)](#bib-dameriaRelationshipResidentsSense2022)

---
# Problem Statement

- **20 residential areas** on **7 districts** on coastal areas are suffering from **tidal flood every year** because the **increase of sea level, coastal abrasion, and land subsidence**

- Semarang sinks up to **15cm/year** due to  the increase of **population density** and **declination of groundwater**

Source: <a name=cite-riaerlaniKetangguhanKotaSemarang2019></a>[Ria Erlani et al. (2019)](#bib-riaerlaniKetangguhanKotaSemarang2019), <a name=cite-widadaDistributianDepthClaySilt2019></a>[Widada et al. (2019)](#bib-widadaDistributianDepthClaySilt2019), <a name=cite-marfaiImpactTidalFlooding2008></a>[Marfai et al. (2008)](#bib-marfaiImpactTidalFlooding2008)

---
#Impact
.panelset[
.panel[.panel-name[Social]

- Around **1660 Ha** area flooded with **72,903 people** living within those area 
- Water supply and electricity cannot be utilized 
- Disruption of household activity 
- High exposure to illness, such as, diarrhea, fever, and malaria 
- Disruption of transportation access
- Causing **isolated residential area**
- Damage of production area (ie fishpond and rice field)
]

<table>
 <thead>
  <tr>
   <th style="text-align:left;"> Land-use </th>
   <th style="text-align:right;"> Estimated Waste Generation (ton) </th>
  </tr>
 </thead>
<tbody>
  <tr>
   <td style="text-align:left;"> Industrial/port area </td>
   <td style="text-align:right;"> 5.9 </td>
  </tr>
  <tr>
   <td style="text-align:left;"> Settlement Area </td>
   <td style="text-align:right;"> 4.9 </td>
  </tr>
  <tr>
   <td style="text-align:left;"> Coastal Area </td>
   <td style="text-align:right;"> 44.1 </td>
  </tr>
  <tr>
   <td style="text-align:left;"> Pond Area </td>
   <td style="text-align:right;"> 15.1 </td>
  </tr>
  <tr>
   <td style="text-align:left;"> Total </td>
   <td style="text-align:right;"> 70.0 </td>
  </tr>
</tbody>
</table>
]

.panel[.panel-name[Economic]
- Spend **£761,236/year** to **handle waste** <a name=cite-rahmaIntroductionStudyTidal2019></a>([Rahma, Maryono, and Widjanarko, 2019](#bib-rahmaIntroductionStudyTidal2019))
- Loss of **£7,714,513/year** because of the **environmental damage**, such as <a name=cite-nugrahaKAJIANPEMANFAATANSRTM2013></a>([Nugraha and Hani’ah, 2013](#bib-nugrahaKAJIANPEMANFAATANSRTM2013)): 
  - Production land damage (fishpond, and rice field)
  - Housing damage
  - Infrastructure damage
]
]

Source: [Rahma, Maryono, and Widjanarko (2019)](#bib-rahmaIntroductionStudyTidal2019)

---
# UN Sustainable Development Goals

**Raise Housing standards, inclusive urbanisation, disaster planning and mitigation**

% urban population in informal settlements

% of cities with civil society planning input

Disaster-related economic, infrastructure and life loss

]

**Build resilience, climate change planning in LDCs **

Sendai Framework implementation

% local government in line with national risk disaster reduction strategies

Contributions, strategies, plans and communications reported to UNFCCC

]
---

# Current Semarang Resilience Strategy
### 3rd Pillar Strategy: Preparedness for Disaster and Disease Outbreak

**A Developing Technology for Disaster Management**

Raise Public Awareness on Disaster-Prone Areas
Explore Alternative Methods to Prevent Dengue Fever
Explore New Technology in Disaster Management

**B Enhancing Capacity of Stakeholders in Disaster and Disease Management**

Replicate Disaster Preparedness Groups in Disaster Prone Areas
Improve A Community- Based Sanitation System
Increase capacity of stakeholder in facing Disaster and Diseases Outbreaks

**C Improve Coordination in Disaster Risk Reduction**
Prepare Participatory Contingency Plan

Source: <a name=cite-semarangcitygovernmentResilientSemarangMoving2016></a>[Semarang City Government (2016)](#bib-semarangcitygovernmentResilientSemarangMoving2016)

---
class: inverse, center, middle

##Unclear what information is used to guide responses.

---
#Compliance to Sendai Framework for Disaster Risk Reduction

**Priority 1: Understanding disaster risk**

Policies and practices for disaster risk management should be based on vulnerability, capacity, exposure of persons and assets, hazard characteristics and the environment

- Knowledge can be used for **pre-disaster risk assessment**, prevention, mitigation and for the development and implementation effective response to disasters.

- To develop, periodically update and disseminate location-based disaster risk information including **risk maps** to decision makers, the public and at-risk communities in an appropriate format by using **geospatial information technology**; and

- To promote real time access to data, make use of space and in situ information including **GIS** and use IT innovations to enhance measurement tools and the **collection, analysis and dissemination of data**.

Source: <a name=cite-unitednationsofficefordisasterriskreductionSendaiFrameworkDisaster2015></a>[United Nations Office for Disaster Risk Reduction (2015)](#bib-unitednationsofficefordisasterriskreductionSendaiFrameworkDisaster2015)
---
#Case Study

### Dar es Salaam, Tanzania
Widespread flooding and informal settlement
50,000 affected, + $780,000 in recovery and emergency response.
Community risk mapping project led by the World Bank and other Tanzanian NGOs

.pull-left[
* Identified substantial geospatial information including landuse, infrastructure and exposure data
* 3.5 million residents in over 228 communities mapped within 3 years.
Flood protective measures could unlock up to $900 million of real estate investments including $200 million of new revenue to the city through the construction of up to 5,900 new housing units.
]

.pull-right[
<img src="img/casestudy_tanzania.jpeg" width="90%" style="display: block; margin: auto 0 auto auto;" />
Image credit: [Ubuntu Times](https://www.ubuntutimes.com/poor-infrastructures-rapid-urban-sprawl-increase-flood-risk-in-tanzanias-largest-city/)
]

---
#Approach

<img src="img/flow_journey.png" width="140%" style="display: block; margin: auto;" />
---
#Approach

---
# Methodology

???

All text under the three question marks are presentation notes; it is hidden from the presentation slides and only available in presenter mode.

---

# Methodology
<img src="img/methodology-flowchart.png" width="130%" style="display: block; margin: auto;" />

---
class: inverse, center, middle
### Example output: Bivariate map

---

# Methodology

'Ensemble-based' Machine Learning approach

1. Using database of historic flood data, split into **training (70%) and testing (30%)** set

1. Build a map (raster) of each hazard input parameter

1. Estimate flood **probability indices** (Frequency Ratio Approach)

1. Apply a **Support Vector Machine (SVM)** model, which will **classify** the input pixels

1. Build a map (raster) of each vulnerability input parameter, classified on a quantile scale

1. **Combine** the hazard and vulnerability maps together to produce a bivariate risk map

1. **Validate** the model using testing set of data

Based off approaches developed by <a name=cite-taubenbockFloodRisksUrbanized2011></a>[Taubenböck et al. (2011)](#bib-taubenbockFloodRisksUrbanized2011), <a name=cite-mojaddadiEnsembleMachinelearningbasedGeospatial2017></a>[Mojaddadi et al. (2017)](#bib-mojaddadiEnsembleMachinelearningbasedGeospatial2017), <a name=cite-tehranyFloodSusceptibilityMapping2014></a>[Tehrany et al. (2014)](#bib-tehranyFloodSusceptibilityMapping2014).

???
Presenter Notes

Refs for this section: Taubenböck et al 2011; Mojadaddi et al. 2017; Tehrany et al. 2014; Flood database: Tellman et al. 2021

Methodology: ‘Ensemble-based’ machine learning using SVM (Support Vector Machine)
Each parameter rendered/built as a raster layer
Frequency Ratio (FR) approach can be used to estimate flood probability indices, using the hazard parameters (the inputs are each classified using quantile breaks and an FR value assigned to each class). This will output which classes of which parameters (e.g. which values of elevation) are at the highest probability of flood. 
Apply an SVM model; essentially a statistical theory to minimise operational risk standard, and performs a binary classification on the input pixels. 
Repeat the process for vulnerability parameters, by classifying the raster cells on a scale (e.g. 0 to 5) of vulnerability depending on the parameter value. 
The Hazard and Vulnerability models can then be multiplied together, and a weighted overlay applied.
A training set taken from historic flood events can be used to validate the built model.

---

# EO Data: Hazard

<table>
 <thead>
  <tr>
   <th style="text-align:left;"> Purpose </th>
   <th style="text-align:left;"> EO data </th>
   <th style="text-align:left;"> Spatial resolution </th>
   <th style="text-align:left;"> Temporal resolution </th>
   <th style="text-align:left;"> Cost </th>
   <th style="text-align:left;"> Processing </th>
  </tr>
 </thead>
<tbody>
  <tr>
   <td style="text-align:left;"> Tidal Range </td>
   <td style="text-align:left;"> BMKG/NOAA Jason-3 </td>
   <td style="text-align:left;"> 3 cm </td>
   <td style="text-align:left;"> near real-time </td>
   <td style="text-align:left;"> Free </td>
   <td style="text-align:left;"> Some basic calculation </td>
  </tr>
  <tr>
   <td style="text-align:left;"> Sea Level Change </td>
   <td style="text-align:left;"> NOAA Jason-3 </td>
   <td style="text-align:left;"> 3 cm </td>
   <td style="text-align:left;"> near real-time </td>
   <td style="text-align:left;"> Free </td>
   <td style="text-align:left;"> Some basic calculation </td>
  </tr>
  <tr>
   <td style="text-align:left;"> Significant Wave Height </td>
   <td style="text-align:left;"> NOAA Jason-3 </td>
   <td style="text-align:left;"> 3 cm </td>
   <td style="text-align:left;"> near real-time </td>
   <td style="text-align:left;"> Free </td>
   <td style="text-align:left;"> NA </td>
  </tr>
  <tr>
   <td style="text-align:left;"> Historic Flood Data </td>
   <td style="text-align:left;"> BMKG/Global Flood Database </td>
   <td style="text-align:left;"> 250 m </td>
   <td style="text-align:left;"> N/A </td>
   <td style="text-align:left;"> Free </td>
   <td style="text-align:left;"> N/A </td>
  </tr>
</tbody>
</table>

Significant wave height (<a name=cite-shahidiTwoNewMethods2020></a>([Shahidi et al., 2020](#bib-shahidiTwoNewMethods2020))): 
$$
E=\frac{1}{8}pgH^2
$$

???

Data pre-processing except Correction and resampling ?  What is our smallest research scale?

Tidal range is the term describing the vertical change between the maximum high tide and the minimum low tide and is related to tidal flooding.(Tidal range is the difference in height between high tide and low tide.)
Indonesia Agency for Meteorology, Climatology, and Geophysics (BMKG)
NOAA Jason-3  is capable of measuring significant wave height (radar altimeter), sigma naught (sigma0), dry and wet troposphere and ionosphere, which can be used to calculate sea surface height, sea surface height anomalies, and total electron content.
A high rate of Sea Level Rise was designated to correspond to high coastal vulnerability and vice versa. 
Significant Wave Height is a replacement of wave energy, it is the average of the highest one-third (33%) of waves (measured from trough to crest) that occur in a given period. Satellite altimetry was the primary data for extracting the significant wave high.

---

# EO Data: Hazard (cont.)

<table>
 <thead>
  <tr>
   <th style="text-align:left;"> Purpose </th>
   <th style="text-align:left;"> EO data </th>
   <th style="text-align:left;"> Spatial resolution </th>
   <th style="text-align:left;"> Temporal resolution </th>
   <th style="text-align:left;"> Cost </th>
   <th style="text-align:left;"> Processing </th>
  </tr>
 </thead>
<tbody>
  <tr>
   <td style="text-align:left;"> Rainfall  </td>
   <td style="text-align:left;"> SM2RAIN-ASCAT </td>
   <td style="text-align:left;"> 1 km </td>
   <td style="text-align:left;"> Daily </td>
   <td style="text-align:left;"> Free </td>
   <td style="text-align:left;"> Apply SM2RAIN algorithm to ASCAT product </td>
  </tr>
  <tr>
   <td style="text-align:left;"> Elevation </td>
   <td style="text-align:left;"> NASADEM Global Elevation Model </td>
   <td style="text-align:left;"> 30m  </td>
   <td style="text-align:left;"> N/A </td>
   <td style="text-align:left;"> Free </td>
   <td style="text-align:left;"> N/A </td>
  </tr>
  <tr>
   <td style="text-align:left;"> Slope </td>
   <td style="text-align:left;"> NASADEM Global Elevation Model </td>
   <td style="text-align:left;"> 30m  </td>
   <td style="text-align:left;"> N/A </td>
   <td style="text-align:left;"> Free </td>
   <td style="text-align:left;"> Slope algorithm in QGIS </td>
  </tr>
  <tr>
   <td style="text-align:left;"> Land Subsidence </td>
   <td style="text-align:left;"> ALOS PALSAR </td>
   <td style="text-align:left;"> 10m </td>
   <td style="text-align:left;"> 11 days </td>
   <td style="text-align:left;"> Free </td>
   <td style="text-align:left;"> Interferometric Synthetic Aperture Radar (InSAR) </td>
  </tr>
</tbody>
</table>

???

Reference Luca Brocca et al 2019
The SM2RAIN–ASCAT data record is obtained from the application of the SM2RAIN algorithm to the ASCAT soil moisture data records H113 and H114 provided by the European Organisation for the Exploitation of Meteorological Satellites (EUMETSAT) https://essd.copernicus.org/articles/11/1583/2019/essd-11-1583-2019.html
The Soil Moisture (SM) product provides an estimate of the water content of the 0-5 cm topsoil layer, expressed in degree of saturation between 0 and 100 [%]. SM processor.
Land Subsidence and InSAR reference https://www.sciencedirect.com/science/article/pii/S0273117718300711

---

# EO Data: Vulnerability

<table>
 <thead>
  <tr>
   <th style="text-align:left;"> Purpose </th>
   <th style="text-align:left;"> EO data </th>
   <th style="text-align:left;"> Spatial resolution </th>
   <th style="text-align:left;"> Temporal resolution </th>
   <th style="text-align:left;"> Cost </th>
   <th style="text-align:left;"> Processing </th>
  </tr>
 </thead>
<tbody>
  <tr>
   <td style="text-align:left;"> Land-use and Land-cover Recognition </td>
   <td style="text-align:left;"> Landsat data (Landsat 8) </td>
   <td style="text-align:left;"> 30 m </td>
   <td style="text-align:left;"> 16 days </td>
   <td style="text-align:left;"> Free </td>
   <td style="text-align:left;"> Supervisory classification, Normalized Difference Built-up Index,Normalized Difference Water Index </td>
  </tr>
  <tr>
   <td style="text-align:left;"> Land-use and Land-cover Recognition </td>
   <td style="text-align:left;"> SAR (Sentinel-2) </td>
   <td style="text-align:left;"> 10-20m </td>
   <td style="text-align:left;"> 10 days </td>
   <td style="text-align:left;"> Free </td>
   <td style="text-align:left;"> Random forest (RF) or multilayer perceptron (MLP) classifiers. </td>
  </tr>
  <tr>
   <td style="text-align:left;"> Building/infrastructure quality </td>
   <td style="text-align:left;"> VHR images (UAV remote sensing) </td>
   <td style="text-align:left;"> 30 cm </td>
   <td style="text-align:left;"> Daily </td>
   <td style="text-align:left;"> $1000 per day </td>
   <td style="text-align:left;"> Texture analysis and supervisory classification </td>
  </tr>
</tbody>
</table>

???

Reference V. H. R. Prudente et al 2020
Normalized Difference Built-up Index, NDBI = (SWIR - NIR) / (SWIR + NIR), NDBI = (Band 6 - Band 5) / (Band 6 + Band 5)  in Landsat 8
Normalized Difference Water Index, NDWI = (Green - NIR) / (Green + NIR), NDWI = (Band 3 – Band 5)/(Band 3 + Band 5)
SAR satellite data can be helpful in LULC mapping in a tropical region with high cloud cover. https://ieeexplore.ieee.org/document/9323404
Unmanned Aerial Vehicle Remote Sensing

---

## Additional data

<table>
 <thead>
  <tr>
   <th style="text-align:left;"> Purpose </th>
   <th style="text-align:left;"> Data </th>
   <th style="text-align:left;"> Scale </th>
   <th style="text-align:left;"> Cost </th>
  </tr>
 </thead>
<tbody>
  <tr>
   <td style="text-align:left;"> Population </td>
   <td style="text-align:left;"> census </td>
   <td style="text-align:left;"> district </td>
   <td style="text-align:left;"> Free </td>
  </tr>
  <tr>
   <td style="text-align:left;"> Economic Activity </td>
   <td style="text-align:left;"> census </td>
   <td style="text-align:left;"> district </td>
   <td style="text-align:left;"> Free </td>
  </tr>
</tbody>
</table>

---

# Recommended Action Plan

???

Climate is changing + informal settlement locations and densities may change rapidly. Maps and plan should be updated annually or every few years 
Data should be available to all relevant stakeholders, so diff departments
aren’t separately collecting data and planning with diff information

Semarang resilience strategy has all these actions that they want to implement in locations that are prone to tidal floods, so with this map they can have a more updated idea of these target areas. One of their key goals is to increase coordination between stakeholders during disasters.

Be specific link it without risk map.

Use this data to make updates to Semarang’s disaster plan:
Develop flood resistant infrastructure, such as roads, in high risk areas.
Update evacuation plan.
Improve drainage systems in high risk areas.
If implementing an early detection warning system, prioritise placement of sensors in high risk areas.
Create a task force to update risk assessment maps annually based on up-to-date data. Update disaster plan accordingly.
Dashboard with risk and vulnerability map with population data; make this available across all relevant sectors - emergency response, health care, city operations.

---

# Resource Allocation

---

# Plan

---

# Project Timeline

---

# Budget Breakdown

---

# Risks and Limitations

- Unexpected employee turnover

- Unexpectedly high costs

- Political disruption

- Disaster during assessment
]

- Consistent data management for smooth transition

- Allow for budget headroom

- Active political monitoring and civil society engagement
]

---

---

# Summary

**Semarang** already had **resilience strategy** to face their annual tidal flood. However, the **strategy is not clear** on what information is used.

Thus, we propose our project to increase stakeholder engagement by **utilising geospatial technology**.

Impact of this project:

- Generate **data driven policy and information**

- **Raise awareness** of the local community

- **Save money** on recovery and emergency response

- And make **Semarang’s strategy** more **tangible and feasible**

---

# References

<a name=bib-dameriaRelationshipResidentsSense2022></a>[Dameria, C., R.
Akbar, P. N. Indradjati, et
al.](#cite-dameriaRelationshipResidentsSense2022) (2022). "The
Relationship between Residents’ Sense of Place and Sustainable Heritage
Behaviour in Semarang Old Town, Indonesia". In: _International Review
for Spatial Planning and Sustainable Development_ 10.1, pp. 24-42.

<a name=bib-marfaiImpactTidalFlooding2008></a>[Marfai, M. A., L. King,
J. Sartohadi, et al.](#cite-marfaiImpactTidalFlooding2008) (2008). "The
Impact of Tidal Flooding on a Coastal Community in Semarang,
Indonesia". In: _The Environmentalist_ 28.3, pp. 237-248.

<a
name=bib-mojaddadiEnsembleMachinelearningbasedGeospatial2017></a>[Mojaddadi,
H., B. Pradhan, H. Nampak, et
al.](#cite-mojaddadiEnsembleMachinelearningbasedGeospatial2017) (2017).
"Ensemble Machine-Learning-Based Geospatial Approach for Flood Risk
Assessment Using Multi-Sensor Remote-Sensing Data and GIS". In:
_Geomatics, Natural Hazards and Risk_ 8.2, pp. 1080-1102.

<a name=bib-nugrahaKAJIANPEMANFAATANSRTM2013></a>[Nugraha, A. L. and H.
Hani’ah](#cite-nugrahaKAJIANPEMANFAATANSRTM2013) (2013). "KAJIAN
PEMANFAATAN DEM SRTM & GOOGLE EARTH UNTUK PARAMETER PENILAIAN POTENSI
KERUGIAN EKONOMI AKIBAT BANJIR ROB". In: _TEKNIK_ 34.3, pp. 202-210.

<a name=bib-rahmaIntroductionStudyTidal2019></a>[Rahma, N. N., M.
Maryono, and W. Widjanarko](#cite-rahmaIntroductionStudyTidal2019)
(2019). "Introduction Study of Tidal Flood Waste Management Cost in
North Semarang Sub-District". In: _E3S Web of Conferences_ 125, p.
07020.

<a name=bib-riaerlaniKetangguhanKotaSemarang2019></a>[Ria Erlani and W.
H. Nugrahandika](#cite-riaerlaniKetangguhanKotaSemarang2019) (2019).
"Ketangguhan Kota Semarang Dalam Menghadapi Bencana Banjir Pasang Air
Laut (Rob)". In: _Journal of Regional and Rural Development Planning_
3.1, p. 47.

---

# References (cont.)

<a
name=bib-semarangcitygovernmentResilientSemarangMoving2016></a>[Semarang
City
Government](#cite-semarangcitygovernmentResilientSemarangMoving2016)
(2016). _Resilient Semarang: Moving Together towards a Resilient
Semarang_. Semarang, Indonesia.

<a name=bib-shahidiTwoNewMethods2020></a>[Shahidi, R. and E. W.
Gill](#cite-shahidiTwoNewMethods2020) (2020). "Two New Methods for the
Extraction of Significant Wave Heights From Received HF-Radar Time
Series". In: _IEEE Geoscience and Remote Sensing Letters_ 17.12, pp.
2070-2074.

<a name=bib-taubenbockFloodRisksUrbanized2011></a>[Taubenböck, H., M.
Wurm, M. Netzband, et al.](#cite-taubenbockFloodRisksUrbanized2011)
(2011). "Flood Risks in Urbanized Areas – Multi-Sensoral Approaches
Using Remotely Sensed Data for Risk Assessment". In: _Natural Hazards
and Earth System Sciences_ 11.2, pp. 431-444.

<a name=bib-tehranyFloodSusceptibilityMapping2014></a>[Tehrany, M., B.
Pradhan, and M. Jebur](#cite-tehranyFloodSusceptibilityMapping2014)
(2014). "Flood Susceptibility Mapping Using a Novel Ensemble
Weights-of-Evidence and Support Vector Machine Models in GIS". In:
_Journal of Hydrology_ 512, pp. 332-343.

<a
name=bib-unitednationsofficefordisasterriskreductionSendaiFrameworkDisaster2015></a>[United
Nations Office for Disaster Risk
Reduction](#cite-unitednationsofficefordisasterriskreductionSendaiFrameworkDisaster2015)
(2015). _Sendai Framework for Disaster Risk Reduction 2015-2030_.
UNDRR.

<a name=bib-widadaDistributianDepthClaySilt2019></a>[Widada, S., M.
Zainuri, G. Yulianto, et
al.](#cite-widadaDistributianDepthClaySilt2019) (2019). "Distributian
of Depth and Clay-Silt to Sand Ratio of Land Subsidence in Coastal
Semarang City by Resistivity Methods". In: _Jurnal Kelautan Tropis_ 22,
p. 63.