Why might incomes of $1 or $2 a day underestimate the value of what is being consumed by Indian households?

Without knowing more about the exact way in which the source got its numbers, it is impossible to be certain about how those figures are potentially wrong. We can, however, at least discuss two possible problems with those numbers.
The first problem is that these numbers might be based on wages paid by the formal economy. In other words, the people who did this study might have used businesses' reports to the government to determine how much people were paid in India. This would be a problem because many people in India work in the informal economy. They do not work for formal companies and their pay is not reported to the government. If the statistics rely only on data from the formal economy, the researchers will miss money people are making informally. Therefore, the figure given here will underestimate how much money people are able to spend on consumption.
The other major way these numbers might underestimate actual consumption is because many people might engage in transactions outside the cash economy. In a rich country, most people are only able to obtain goods or services by paying for them with money. In India, many people might be engaging in barter and/or subsistence farming. If I only have $2 per day to live on, but I grow much of my own food, the $2 figure will underestimate how much I am able to eat.  Similarly, if I take my excess crops and trade them with other people for goods or services (instead of selling them for money), I will be able to consume more things without having any actual money. Thus, if people are engaging in subsistence farming and/ or a barter economy, the $2 figure could underestimate how much they are actually able to consume.
Thus, these numbers might be misleading. They might not take into account money that people earn outside official employment and they might fail to account for goods and services people make for themselves or for which they barter.
https://www.economist.com/asia/2013/10/03/hidden-value