Impact of Past Rainfall Events on the Urban Transport Sector of the Mumbai Metropolitan Region: Current and Future Projections Under BAU Scenario

Abstract

This paper analyses the impact of antecedent rainfall events on the urban transport of the Mumbai Metropolitan Region (MMR), Maharashtra, India. The impacts are analysed in terms of cancelled trips, passenger and vehicle kilometres travelled. We utilized a high flood level map for rainfall events between 2005 and 2007 to prepare a flooded transport network for MMR from 2005 to 2050. This flooded network was modelled in a travel demand model for MMR, restricting the speed and public transport access links based on the flood depth. The results show that maximum cancelled vehicle trips in case of floods would be from private mode (~ 60%) in 2050. The cancelled passenger trips from the metro and suburban rail would be ~ 52% due to both rainfall events. This decrease in the trips would contribute to the reduction in passenger and vehicle activity by an average of ~ 45% and 75% in 2050. The analysis of this study will be beneficial for policymakers to implement various policies and remedial measures towards reducing the effect of such rainfall events in the future.

Appendix 1: Data Inputs, assumptions and Methodology Used in the Formulation of Travel Demand Model

Data Inputs

The travel demand is based on the comprehensive four-stage process and we utilized the inputs from comprehensive transport study (CTS) [33]. We utilize the following inputs from CTS [33]:

• We utilized 1037 traffic analysis zones in Mumbai Metropolitan Region (MMR) among which seven were external zones to capture external vehicle trips to or from MMR.

• We utilized planning parameters for the travel demand model in terms of population and employment. Population (Pop) was segregated into four different categories: resident workers employed office-type jobs (RWF), resident workers employed industrial type jobs (RWI), resident workers employed in other jobs (RWO), and resident students (RS). Employment (EBZ) was segregated into three categories: office-type jobs (OJ), industrial type jobs (IJ) and other type jobs (OJ). We utilized six travel purposes; home-based work purpose employed in office-type jobs (HWF), home-based work purpose employed in industrial type jobs (HWI), home-based work employed in other type jobs (HWO), home-based education (HBE), home-based other (HBO) and non-home based (NHB).

• The base year transport network for MMR included road and rail networks along with their attributes and speed and delay characteristics. The commercial and external vehicle demand, passenger car unit (PCU) factors and vehicle occupancy rates were taken from CTS for MMR.

The horizon year travel demand model was based on the following assumptions:

Horizon Transport Network

We incorporated the proposed infrastructure in the base year network. We added the two new modes, i.e., metro and monorail, along with their routes [36, 37] in base year transport network. Similarly, we added three major highway corridors, i.e., Eastern freeway [38], Mumbai Trans Harbor link [39], and Multimodal Corridor from Virar to Alibaug and the proposed extension of suburban rail network [40].

Planning Parameters

We assumed a combined annual growth rate (CAGR) of 1.9%, 2.3%, and 1.9% for MMR’s population in 2001–21, 2021–31, and 2031–50, respectively [33]. For the period 2041–50, CAGR for the population has been assumed the same as 2031–41. Similarly, we assume a CAGR of 1.77%, 3.28%, 2.42% and 2.19% for MMR’s employment 2005–11, 2011–16, 2016–21 and 2021–50.

External and Commercial Vehicle Demand

We assume a CAGR of 3% and 2.35% for internal and external commercial vehicles, respectively. For, external vehicles, we assume a CAGR of 10.47% and 5.48% for personalized vehicles and buses, respectively [33].

Methodology

Trip Generation

Multiple linear regression (MLR) was used to obtain trip productions and attractions for each travel purpose in MMR while utilizing planning parameters as exogenous variables. The production factors and attraction factors were taken from CTS [33].

Trip Distribution

We used gravity model in this stage and estimated the friction factors from Tanner’s function:

$${T}_{ij}= {a}_{i}{P}_{i}{b}_{j}{A}_{j}{f}_{ij}$$

where, ${a}_{i}=\frac{1}{\sum_{j=1}^{n}{b}_{j}{A}_{j}{f}_{ij}}$, ${b}_{j}=\frac{1}{\sum_{i=1}^{n}{a}_{i}{P}_{i}{f}_{ij}}$, T_ij = number of trips from zone I to zone j; P_i = number of trip productions in zone i; A_j = number of trip attractions in zone j; f_ij = “friction factor” or “deterrence function” relating the spatial separation between zones i and j; n = total number of zones.

Modal Split

We utilized the multinomial logit (MNL) model while using utility equations obtained from CTS [33]

$${P}_{i}= \frac{{e}^{{U}_{i}}}{\sum_{i=1}^{n}{e}^{{U}_{i}}}$$

where P_i = probability of mode i to be chosen; U_i = Utility equation of mode i, n = total no. of mode choices.

The mode share obtained in the base year and horizon years is shown in Fig. 7.

Fig. 7

Mode share of internal vehicles of MMR in the base year and BAU scenario

Traffic Assignment

We used a multimodal capacity restrained equilibrium technique to assign traffic to the transport network. The public transport assignment followed by highway assignment (private, commercial, and external vehicles).

Model Validation: Vehicle Flows, Passenger Flows and Mode Share

Tables 2 and 3, Fig. 8 show the validation results for Vehicle flows (PCU), Rail passenger flows, and mode share, respectively. As the difference was in the acceptable limit of ± 30%, the model was considered as validated.

Table 2 Private, commercial, bus and IPT vehicle flows (PCU)

Table 3 Rail passenger flows

Fig. 8

Comparison of estimated and observed mode share of base year