Agree that it makes sense to hike rewards on Base given TVL growth and the comparative under-rewarding relative to mainnet. @prevert raises a good point on the cost-effectiveness of this.
In terms of modeling and longer-term strategic impact, definitely in favour of a more granular reward models per asset class rather than a broad strokes approach. Gradually scaling rewards while collecting relevant data is a no brainer to help the DAO allocate rewards in a more capital-efficient way. Over time, creating a more dynamic model based on asset sensitivity to rewards, adjusted as the DAO scales across chains, will also be important to optimize cross-network and asset-class reward distributions as the protocol matures.
Also, while TVL is an important metric, it may be worth exploring other data points related to user retention rates, liquidity utilisation, volume, avg. loan duration etc which could help provide a more holistic view of liquidity health when adjusting rewards.