Clear ideas for complex systems
Latest posts
Energy and Climate Systems

The Duck Curve: Solar, Ramp Constraints and Grid Flexibility as a System

The duck curve as a feedback system: why solar reshapes net load, how ramp and minimum generation bind, and which levers flatten the evening climb.

Table of ContentsNavigate this article

    Last updated: 31 October 2026.

    The duck curve is the shape of net electricity demand on a grid with a lot of solar power: demand minus solar output falls through the middle of the day, then climbs steeply as the sun sets and people come home. California’s grid operator popularised the chart in the early 2010s as a warning that rising solar capacity changes the shape of the problem before it changes the amount of energy needed. For systems thinkers, it is a clean example of a balancing system whose solution to one constraint (cheap, clean daytime energy) creates a new constraint somewhere else (a steep evening ramp), with flexibility, storage and market rules acting as the delayed corrections. This article explains the mechanism with an illustrative model, then looks at the feedbacks and levers.

    Key takeaways

    • Net load is demand minus variable generation. Solar pushes it down at midday and leaves the evening peak untouched, so the ramp the rest of the fleet must deliver gets steeper.
    • The binding constraints are not total energy but ramp rate, minimum generation and timing.
    • Midday surplus brings low or negative prices and curtailment, which weakens the economics of additional solar: a balancing feedback.
    • Storage, flexible demand, geographic diversity and market design shift energy in time and flatten the curve. Each has limits and delays.

    How the duck forms

    Grid operators must balance supply and demand at every instant. The relevant quantity for conventional plants is not total demand but net load: demand minus the output of wind and solar, which are dispatched first because their marginal cost is near zero. On a sunny day, solar output rises from sunrise, peaks near noon and drops to zero at sunset. Household demand, in contrast, tends to rise in the late afternoon and evening, when people return home, cook and switch on lighting and cooling. Subtracting solar from demand therefore carves out a belly at midday and leaves a tall neck at dusk.

    Illustrative net load across a day with high solarA line chart over 24 hours. Total demand is a smooth curve with an evening bump. Solar output is a bell shape around midday. Net load, demand minus solar, dips at midday and then climbs steeply in the evening, forming the duck-shaped curve. Numbers are illustrative.00:0006:0012:0018:0024:000102030Hour of dayPower (GW, illustrative)Demand (dashed)Solar outputNet loadEvening ramp:+11 GW in 3 hours

    Figure 1. Illustrative net load on a solar-heavy day. Demand (solid) is reduced by solar output at midday, producing a deep trough around noon and a steep three-hour rise in the evening. Numbers are generated by a simple model for illustration and do not describe any specific grid.

    In my illustrative model (a stylised daily demand curve with a bell-shaped solar profile), net load reaches a minimum of about 12 GW at noon, and the steepest three-hour increase is about 11 GW, starting around 15:00. The absolute values depend on the assumptions, but the pattern is the point: the evening ramp is set by how fast solar falls away while demand rises, and it grows with the amount of solar installed.

    Three constraints that bite before energy runs out

    1. Ramping

    Dispatchable plants can only change output at a limited rate. When net load climbs by many gigawatts within a few hours, enough units must be online, positioned and able to ramp. Gas turbines, hydro and batteries can respond quickly, while large thermal units are slow. The system needs fast flexibility at precisely the time when solar is disappearing.

    2. Minimum generation

    Some plants must stay on because shutting down and restarting is slow, costly or restricted by reliability rules. If net load at noon falls below the combined minimum of the must-run units, there is a surplus. The options are to curtail solar, export, store or pay someone to take the power. This is why midday prices can fall to zero or below.

    3. Timing mismatch

    Energy is abundant when demand is soft and scarce when demand is high. Adding more solar raises midday abundance without helping the evening. The marginal value of the next unit of solar capacity falls, an effect usually called solar cannibalisation: the more solar there is, the lower the price it captures.

    The feedback structure

    The duck curve shows two linked loops:

    • B1: cannibalisation (balancing). More solar leads to a deeper midday trough, which leads to lower prices and more curtailment, which reduces returns on new solar and slows additions. The system pushes back against its own growth.
    • B2: flexibility response (balancing, delayed). A steeper ramp and price spread raise the reward for storage, flexible demand and fast plants. As they are built, the ramp is absorbed and the curve flattens. The delay comes from planning, procurement and construction.

    Between the two, the system tends to oscillate: periods where flexibility lags and operators curtail renewable energy, then periods where storage catches up. This lag between a problem and its corrective investment is the standard structure explained in delays in systems thinking. Policy is a third actor: subsidy design, interconnection queues and market rules influence how quickly B2 operates relative to B1.

    Efficiency versus resilience

    A grid optimised for lowest cost at average conditions keeps little spare ramping capacity. The duck curve exposes this: efficiency in a world of stable net load becomes fragility when the net load shape changes. The trade-off is the subject of resilience versus efficiency. Flexibility is a form of redundancy, and like all redundancy it appears wasteful until it is needed.

    There is also a path-dependence angle. Existing gas fleets, contracts and market rules built for a different load shape can slow adaptation, a theme that links to carbon lock-in. New demand adds another layer: large data centres run near-constant load, which partly fills the midday trough but adds to the evening and night base; see AI data centres and the power grid for that interaction.

    Levers that flatten the duck

    Lever How it helps Limits
    Battery storage Charges at midday, discharges at the evening peak Duration, cost, siting and interconnection delays
    Flexible demand Shifts loads (EV charging, water heating, industry) to midday Needs price signals, controls and customer participation
    Geographic and technology diversity Wind and west-facing solar shift the shape; regional trade moves surplus Transmission capacity and market integration
    Fast-ramping plants and hydro Cover the evening climb Fuel, emissions and availability
    Market and tariff design Rewards flexibility and time-shifting Regulatory lag and political acceptance

    The U.S. Energy Information Administration reported that battery storage capacity in California grew from about 0.2 GW in 2018 to about 4.9 GW by April 2023, a rapid build-out consistent with the B2 response. Check current figures from the EIA and the grid operator, since the numbers keep changing quickly.

    Common misconceptions

    • “The duck curve means solar does not work.” It means solar changes the shape of the residual problem. Grids around the world manage much higher solar shares by adding flexibility.
    • “It is only a California issue.” Any grid with high solar penetration and an evening peak faces a version of it, with different details by region and season.
    • “Storage alone solves it.” Storage helps for hours, not for multi-day shortfalls, and it is one of several levers.
    • “Curtailment is waste, so build less solar.” Some curtailment can be economically optimal. The right question is the cost of the last megawatt-hour compared with alternatives.

    Flexibility, storage and local conditions

    Solar output is predictable in broad daily shape, but the system response depends on what else is operating at the same hour. Flexible demand can move some consumption into sunny periods: water heating, industrial processes, vehicle charging and thermal storage may all help when schedules and infrastructure allow. The useful question is not simply whether demand can shift, but how much can move, for how long, and whether the shifted load creates a new evening peak.

    Storage can absorb midday surplus and return energy later, but its contribution depends on power, energy capacity, efficiency, state of charge and duration. A battery sized for a short evening ramp does not solve a multi-day low-renewable event. Pumped hydro, thermal storage, interconnection and dispatchable generation provide different services and constraints. Planning should compare the duration and location of the need with the duration and location of each resource.

    Geography and the rest of the generation mix matter too. A regional curve can conceal local transmission bottlenecks, while a system with hydro, wind or flexible thermal units may respond differently from one dominated by inflexible generation. Weather correlation also matters: neighboring regions may have surplus at the same time, reducing the value of trade. A practical assessment uses hourly data across seasons, tests transmission limits and distinguishes energy adequacy from ramping capability.

    These distinctions keep the duck curve useful as a diagnostic rather than a forecast by itself. Track midday prices, curtailment, net-load ramps, reserve needs and evening capacity together. If solar additions deepen the midday trough but flexibility grows as well, the system can accommodate more clean energy without treating every visible dip as a reliability crisis.

    Interpret the curve with operational data

    Use hourly net load alongside generation, curtailment and reserve data to locate the actual constraint. Compare seasons and weather years, and test what happens when transmission or neighboring supply is limited. This grounds the metaphor in system operation and helps distinguish a steep ramp from a genuine shortage of dependable capacity.

    Frequently asked questions

    What is the duck curve in simple terms?

    It is the daily graph of net load, demand minus solar and wind, which dips in the middle of the day and rises sharply in the evening, forming a duck-like shape.

    Why does a deeper duck curve make new solar less valuable?

    Because extra solar adds energy when there is already a surplus, pushing midday prices down and increasing curtailment, so each additional unit earns less.

    How do batteries change the duck curve?

    They absorb midday surplus and release it in the evening, raising the midday net load and lowering the evening ramp. The effect depends on their power and duration.

    Does the duck curve apply to the whole year?

    The shape varies with season. It is typically most pronounced on mild, sunny days with low demand, and differs on hot summer afternoons or in winter.

    Conclusion

    The duck curve is an example of success creating its own constraint. Cheap midday solar is valuable, and it moves the problem to ramp rate, minimum generation and timing, where the correction depends on investment that lags the cause. Treating it as a feedback structure explains the price signals, the oscillations in curtailment and the importance of flexibility as a form of resilience. For the international negotiation context for the energy transition, see climate negotiations as a system, and see the glossary entry on the duck curve.

    Sources and further reading

    • U.S. Energy Information Administration, article on California battery storage growth (Today in Energy).
    • California ISO, “What the duck curve tells us about managing a green grid”, Fast Facts, 2016.
    • Denholm, P., O’Connell, M., Brinkman, G. and Jorgenson, J., “Overgeneration from Solar Energy in California: A Field Guide to the Duck Chart”, NREL Technical Report NREL/TP-6A20-65023, 2015.
    About the publisher

    Systems Thinking Hub

    We are experts in delivering clear, practical knowledge about systems thinking, complexity, and better decision-making.

    Discussion

    Join the discussion

    Your email address will not be published. Required fields are marked.