Digital Innovation in Water Scarcity

Coimbatore, India

Research

The project is an intervention in the current water crisis in Coimbatore, Southern India. The aim of the research project is to be an urgent response to strained and failing strategies of water conversation policy and practice. The project will have immediate and measurable impact and will establish a sustainable legacy of data for ongoing and subsequent development-led research in the region and beyond. The research will use digital humanities methods to assemble data related to imaginaries, regimes and economies of water bodies over a 200-year period. Through the combination of transparent and reflexive practice, the research aims to establish this database as a trusted resource for a range of stakeholders. The research will employ GIS-based strategies to challenge the disambiguation of land and water that was created by colonial cartographies in the nineteenth century and to 'return' archived data to localities and to communities.

The project will generate innovative, integrated and sustainable conservation strategies.

The first stage will create data sets about Coimbatore District from the wealth of textual and cartographic materials about the Madras Presidency and Tamil Nadu. This data will be assembled from a expansive range of pre-digitised texts and the corpus of maps held by the National Library of Scotland. These maps will be digitised and georeferenced for the project.

Digital Humanities methods of mining and layering will be used to collect quantitative, qualitative and spatial data about water in any form: aquifers, tanks, riverine systems, domestic water supplies, industrial exploitation, hydraulic interventions (dams, embankments, canals), rainfall and puddles. This data will reveal patterns of historic and contemporary water security and water scarcity in Coimbatore on an unprecedented scale. This database will be developed and refined across the duration of the project and will be made available in a user-friendly format as a free, open-access resource.

The second aspect of the project is intensive fieldwork in the Coimbatore. A series of local interviews and interactions will discuss the aims of the project with local activists, scholars and residents. A key aspect of this work will be to garner information about living, local narratives of water scarcity. The water crisis is universally acknowledged but narratives are characterised, and compromised, by fragmentation. The integrity of the project's intervention will depend upon our ability to engage with those fragmentations and to establish trust with local stakeholders.

The third aspect of the project will combine the assembled data with the understandings garnered through local interactions to create interactive, culturally resonant digital visualisations designed to disrupt catastrophic patterns of water exploitation in the region. These visualisations will combine and manipulate the data collected in order to create new narratives about the history of water. The visualisations will aim to change the narratives about scarcity, the landscape and rights. The visualisations will challenge the tacit spatial and temporal assumptions that define the categories that separate water from land. 

One of the broader outcomes of the project is to build and consolidate networks of scholars and environmentalists within South Asia and internationally who will draw on the water data assembled and engage with the visualisations created. The digitisation and dissemination of archived materials, and data derived from them, raises a number of difficult questions about the subsequent use of that data in the contexts of informal and profoundly unequal terms of property and resource entitlement. A key aim of the project is to conceptualise a strategy to nuance a methodological commitment to open data for data-led projects in the Global South. This approach will engage with analogous data interventions elsewhere and refine an approach to sensitive data.

Methods

The PI and RA will mine the data relating to imaginaries, regimes and economies of water bodies and identify the vanished and shrinking water bodies through various digitized and non-digitized cartographies, revenue, military, famine records, archival, artistic, archaeological, agrarian irrigations, sketches, photos and statistical handbooks. The textual corpus will be largely pre-digitized through targeted research will be carried out in the India Office Collections of the British Library and Tamil Nadu State archives. Text mining and geo referencing tools will be employed for collecting data. For text mining, we will deploy Natural Language Processing (NLP) and corpus linguistics tools. NLP and corpus linguistic tools, such as Named Entity Recognition, Keywords in the Context and Collocation, will be used to extract both statistical and textual data. Data to be collected for the region includes: the number, volume and width of water bodies, the position and changing condition of water bodies in urban, semi-urban and rural contexts, the number of dry and perennial water resources, data relating to extreme seasonal changes (drought and flood), toponymies relating, and dislocating, water bodies from place. 

Geotagging spatial entities within the corpus will face the challenge of non-standard transliteration. The project will benefit from and extend the historic gazetteer created by the PI as part of the Pelagios Commons project (May to November 2019). Once the data is assembled, key research questions will include:  How can local and regional hydraulic transformations in the region be spatially associated?  When, and on what terms, does over-exploitation and scarcity enter various types of texts within the corpus?  How many water bodies are disappeared and shrinking in the places of urban, semiurban and rural?  What are the main extreme seasons of droughts and floods that affect the region in terms water quality and quantity? (for example, neglected water bodies during the drought season are occupied by the weeds, human occupation and sewages which lead to secondary flooding and water pollution The project will use the extensive map collection of the National Library of Scotland (NLS), supplemented by open-source maps from a number of other archives. More than 100 maps of ½ and 1 inches are available particularly on Coimbatore district under Madras Presidency from the year of 1856 to c.1949. At the NLS, hundreds of survey maps include information about geography, geology, topography, water resources, administration, and infrastructural resources such as railway, roads, telegraphs. The British Coimbatore district map covers 10 talukas (tehsil): Bhavani, Coimbatore, Dharapuram, Erode, Karur, Kollegal, Palladam, Pollachi, Satyamangalam, Udamalpet. Preliminary sampling of the maps available at the NLS have demonstrated the wealth of data available about canals, lakes, tanks, rivers, deserted villages, villages, railroads, towns, cities, telegraphs. These maps will be digitised and georeferenced by the NLS. This work should be completed by the end of February 2020. The maps will be interrogated in two ways: 1. Vectorising to extract the spatial, qualitative and quantitative information they contain, and 2. To interrogate the conceptual dogmas of resource classification they embody and, by cross-referencing their graphic form with textual and ethnographic information, to discover their omissions and elisions. We will deploy Historical GIS techniques by superimposing the old maps on the current satellite imagery through georefencing (through carefully adding more ‘control points’ from the old maps to satellite map). The Transparency tool in ArcGIS will allow us to view, compare and analyze changing cartographic representations of water in Coimbatore and connect those representations to the data mined from a range of textual sources.