{"id":"9bfb2a50-7752-4287-afcc-b9a152b49c54","arxiv_id":"2501.13868","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Existing US rooftop solar installations are disproportionately located in lower-carbon-offset areas, and rerouting new installations could improve carbon reductions by up to 39.8% while keeping most energy output.","lead":"This paper connects where rooftop solar panels are installed in the US to how much carbon they actually offset, finding that areas that could offset the most carbon have the fewest panels. It then simulates smarter placement strategies and shows that a balanced approach could cut up to 39.8% more carbon than continuing today's pattern.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Round Robin simulation may not cap panels per ZIP at rooftop capacity, so the claimed 39.8% carbon improvement and 94.6% energy fraction could describe infeasible deployments.","rationale":"The reader's weakest_assumption is the disclosed carbon-offset timing limitation, which is a real data-quality caveat but is explicitly acknowledged in Section 3.2.1 and is unlikely to reverse the paper's qualitative finding that high-carbon-offset regions are underserved. My principal concern is instead an unacknowledged modeling gap: the Round Robin simulation, which directly produces the headline 39.8% and 94.6% figures, is not described as enforcing per-ZIP rooftop capacity limits, unlike the other strategies that explicitly exhaust 'available rooftop space.' Because the simulation places up to 1.8 million panels across about ten thousand ZIP codes, the algorithm necessarily revisits ZIPs; without a capacity cap, the results describe an impossible deployment. This is exactly the kind of hidden assumption that can change the relative ranking of strategies and thus the central quantitative claim. The proposed code inspection and re-run with capacity caps would settle whether the reported improvements survive. Since the issue is fixable and the paper's qualitative conclusions may still hold, I keep the reader's CONDITIONAL verdict rather than moving to REJECT. I mark disagreement with the reader's weakest_assumption because my load-bearing concern is the capacity constraint, not the carbon-offset timing, though the reader's rationale did mention capacity as a secondary issue.","tokens_in":18110,"tokens_out":12174,"duration_ms":111754,"concrete_test":"Inspect SunSight's visualization/projections_util.py to determine whether each ZIP's panel count is capped by an estimated maximum install count (e.g., Sunroof viable rooftops or the inverse of realized potential times existing installs). If no cap exists, re-run the Round Robin and Status Quo simulations with per-ZIP caps equal to that maximum, then recompute carbon-offset and energy ratios at N=1.8M and at the N value that produces the reported 39.8% improvement. If either ratio drops materially (e.g., more than 5 percentage points), the headline claim requires qualification.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Section 4 (Strategy 5) defines the 'Round Robin' policy as adding panels by cycling through the energy, carbon, and equity sorted lists, but, unlike the energy- and carbon-efficient strategies, which explicitly install 'until the available rooftop space is exhausted,' it never states that per-ZIP installations are capped at the number of viable rooftops. With N up to 1.8M panels across only ~10,559 ZIP codes, the algorithm must make multiple passes through the sorted lists; if capacity is not enforced, top-ranked ZIPs can receive more panels than their roofs can hold. This is not a minor implementation detail: the claimed improvements are ratios between simulated strategies, and an uncapped Round Robin can overstate carbon offsets by continuing to place panels in high-offset ZIPs that would actually be saturated, while an uncapped Status Quo can overstate how much deployment continues in already-saturated regions. The paper never describes how 'available rooftop space' is quantified or applied in the Round Robin simulation, nor does it report whether SunSight enforces it. The absence of this cap makes the headline 39.8% carbon-efficiency gain and the 94.6% energy preservation potentially infeasible as stated.","agreement_with_reader":"disagree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper combines Google Project Sunroof, ACS5, Ember, and MEDSL data to study the alignment between residential rooftop PV siting and carbon offset potential in the United States. It reports that ZIP codes with higher carbon offset potential tend to have fewer existing installations, with disparities along racial, income, and political lines. It then proposes and simulates five siting strategies, including a multi-objective 'Round Robin' strategy, and claims that this strategy achieves 94.6% of the energy generation of the status quo while improving carbon offsets by 39.8%. The authors release the SunSight toolkit comprising the cleaned data, analysis scripts, and simulation framework. The paper's central contribution is the quantification of a 'hidden carbon cost' of inequitable siting and the demonstration that simple alternative strategies could improve the climate impact of future residential solar deployment.","tokens_in":18293,"tokens_out":4543,"duration_ms":39378,"significance":"If the results hold, the paper makes a timely and policy-relevant contribution by connecting distributional inequity in rooftop solar to carbon efficiency, an angle that prior work has largely treated separately. The observed demographic patterns are broadly consistent with earlier studies (e.g., Sunter et al.), and the addition of a carbon-offset dimension is a genuine extension. The public release of SunSight is a practical strength, as is the use of externally sourced data and the absence of fitted parameters in the core analysis, which makes the simulation pipeline transparent and reproducible. However, the headline 39.8% improvement and 94.6% energy-preservation figures depend on simulation assumptions that are not fully specified, and the realized-potential metric that underpins the equity analysis is ambiguously defined. These issues affect the load-bearing quantitative claims and require strengthening before the results can be considered robust.","major_comments":[{"comment":"The Round Robin simulation, and also the Status Quo simulation in Strategy 1, are not described as enforcing any cap on the number of panels per ZIP code. Unlike the energy- and carbon-efficient strategies, which explicitly install panels 'until the available rooftop space is exhausted,' the Round Robin description says only that panels are added by cycling through sorted lists. With N up to 1.8 million panels across roughly 10,559 ZIP codes, an uncapped procedure would place implausibly large numbers of panels in a few high-ranked ZIPs, and the claimed 39.8% carbon improvement and 94.6% energy fraction could be artifacts of deployments that exceed rooftop capacity. The paper must specify how 'available rooftop space' is quantified (e.g., from Project Sunroof's viable-roof estimates) and confirm that all simulated strategies respect this capacity constraint, then re-run the simulations with the constraint enforced and report whether the headline ratios persist.","section":"Section 4, Strategy 5"},{"comment":"The definition of Realized Potential in Section 3.1 says it is 'the number of existing solar panel installations as a percentage of the potential solar PV installations,' but the procedure in Section 3.2.1 uses percent-covered to scale existing installations (multiplying by 1/percent-covered) to 'estimate the total.' This scaling estimates the total number of existing installations, not the number of potential installations, so it is unclear how the denominator of realized potential is obtained. If percent-covered refers to the fraction of buildings analyzed, the scaling only corrects for incomplete coverage of the existing-install count, not for the total viable-roof count. This ambiguity affects the realized-potential statistics reported in Table 2 and Figure 4, which are central to the paper's equity claims, and could change the demographic comparisons if percent-covered varies systematically by income or race. The authors should clarify the exact construction of realized potential and justify the scaling assumption.","section":"Section 3.2.1"},{"comment":"The carbon offset potential used throughout the paper is based on eGRID non-baseload emission rates and does not account for the timing mismatch between PV generation and grid demand (the duck curve). The paper acknowledges this limitation but does not test its impact on the results. Because the core finding is that carbon offset potential is geographically misaligned with existing installations, a systematic regional bias in the emission-rate assumption could affect both the observed correlation and the simulated strategy comparisons. A sensitivity analysis using alternative marginal emission factors, or a time-matched generation and dispatch model for a subset of regions, would strengthen the central claim. If such data are unavailable, the paper should at least quantify the plausible range of error and discuss how it might alter the 39.8% improvement figure.","section":"Section 3.2.1, drawback (4)"},{"comment":"The abstract states that neighborhoods with relatively higher Black population have '7.4% higher carbon offset potential than average but 36.7% fewer installations,' while Table 2 reports +6.9% carbon offset potential and -34.0% realized potential for the same group. These are different numbers for what appears to be the same comparison. The discrepancy needs to be reconciled or the text should clarify that different subgroup definitions are used. Since these quantitative claims are a prominent part of the paper's motivation, an unexplained mismatch undermines the reliability of the reported statistics.","section":"Abstract vs. Table 2"}],"minor_comments":[{"comment":"The text introducing Figure 7 says 'we plot projections of the additional energy generation (in kWh)' but the figure and caption describe carbon emissions offset; this should be corrected to 'carbon offset.'","section":"Section 6.2"},{"comment":"The Round Robin description says the strategy cycles through 'energy, carbon, equity' lists, but Section 4 defines two separate equity strategies (Black population proportion and low median income). It is unclear whether the Round Robin uses one combined equity list or cycles through four lists (energy, carbon, Black, income). Clarify the exact construction.","section":"Section 4, Strategy 5"},{"comment":"The caption says Round Robin is a 'Round Robin of each of the other strategies,' which suggests it combines all objectives, but the body text only mentions energy, carbon, and equity. Align the caption and text.","section":"Figure 8 caption"},{"comment":"The paper refers to 'Google's Project Sunroof API' in Section 1.1 but the methodology in Section 3.2.1 describes retrieval of the published dataset; verify the terminology and data-access method for consistency.","section":"Section 3.2.1"}],"recommendation":"major_revision","confidential_remarks":"The paper's main quantitative claims are conditional on simulation details that are not currently specified, particularly the enforcement of rooftop capacity in the Round Robin and Status Quo strategies. The toolkit's public availability is a strength, and the authors could likely address the capacity concern by adding explicit caps and re-running the simulations. I would also check the actual SunSight code (if available at the linked repository) to see whether capacity constraints are already implemented; if they are, the manuscript's omission becomes a presentation issue rather than a correctness issue. The abstract/Table 2 discrepancy and the realized-potential definitional ambiguity should be fixed before acceptance."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The honest summary is that this paper's observational core is real and worth taking seriously, but its headline number is not yet at that same standard. What is genuinely new: it connects the known demographic inequities in US rooftop solar to carbon offset potential, at ZIP-code scale, using Project Sunroof plus ACS, Ember, and MEDSL. The finding that areas with higher carbon offset potential have fewer existing installations—e.g., higher-Black and lower-income ZIPs with 7-15% more offset potential and 34-47% fewer installations—is consistent with prior equity literature and looks like a robust pattern, not an artifact. That alone is a useful contribution. The authors also deserve credit for shipping the SunSight toolkit and for listing the Project Sunroof limitations (CV errors, incomplete coverage, duck curve) rather than burying them.\n\nThe soft spots are in the simulation layer. The stress-test concern about the Round Robin strategy is legitimate: unlike the energy- and carbon-only strategies, which are said to exhaust available rooftop space before moving on, the Round Robin description never mentions a per-ZIP capacity cap. With N up to 1.8M panels across ~10.5k ZIPs, the algorithm makes multiple passes; if capacity is not enforced, top-ranked ZIPs can receive more panels than roofs can hold, and the 39.8% carbon improvement plus 94.6% energy fraction would describe infeasible deployments. The public code can settle this, but the paper as written does not. Second, carbon offset potential uses eGRID non-baseload rates and ignores the timing mismatch between PV generation and demand; the authors flag this, but if the bias is geographically correlated with demographics, it could affect both the observed disparity and the simulated gains. Third, there is no uncertainty quantification: no sensitivity analysis over eGRID vintage, Sunroof CV error, or the crude percent-covered scaling used to define realized potential. These are fixable and do not undermine the direction of the argument, but they do mean the exact 39.8% and 94.6% numbers should not be used in policy yet.\n\nWho benefits: energy systems researchers, environmental justice scholars, and e-Energy readers who want a concrete, data-grounded argument that siting policy can improve carbon outcomes while reducing inequity. It deserves a serious referee; I would recommend conditional acceptance, with the capacity-cap check and a robustness section as the conditions.","headline":"The observational link between solar siting inequity and carbon offset potential is real and worth citing; the 39.8% simulation headline needs a capacity-cap check and uncertainty analysis before being used in policy.","tokens_in":18827,"tokens_out":2456,"would_cite":true,"duration_ms":23666,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"Residential solar in the US is concentrated where it offsets the least carbon, and a simple multi-objective siting rule could improve its climate benefit by up to 39.8% while keeping 94.6% of the energy gain.","keywords":["residential solar photovoltaics","carbon offset potential","solar siting strategies","energy equity","grid carbon intensity","Project Sunroof","multi-objective optimization","decarbonization"],"falsifier":"Recompute per-ZIP carbon offset potential using hourly marginal grid emission rates—the emissions of the actual plants that would ramp down when solar generates—instead of yearly average non-baseload rates, and rerun the siting simulations; if the resulting carbon offset rankings stop showing high-offset, low-installation ZIPs, or if the Round Robin strategy no longer beats the status quo in avoided carbon, the central claim is refuted.","tokens_in":17834,"feed_emoji":"☀️","tokens_out":7612,"duration_ms":65755,"temperature":0.7,"pith_summary":"Residential solar panels in the US are not being built where they would do the most for the climate. Combining Project Sunroof's per-panel carbon offset potential with census and voting data, the paper finds a systematic inversion: neighborhoods with higher shares of Black residents have about 7.4% higher carbon offset potential than average but 36.7% fewer installations, and lower-income neighborhoods have 14.7% higher potential but 47% fewer installations. Because installations are concentrated in high-sun, low-carbon-offset regions, the status quo trajectory wastes a large share of the potential climate benefit of each new panel. The paper proposes siting strategies, including a multi-objective 'Round Robin' rule that cycles among energy, carbon, and equity priorities, and simulates that it would avoid up to 39.8% more carbon than continuing current trends while still achieving 94.6% of the status quo's added electricity generation. If true, modest changes in where solar is encouraged could make residential PV substantially more effective as a decarbonization tool while narrowing demographic inequities.","feed_headline":"Round-robin solar siting cuts carbon 39.8% at little energy cost","feed_subtitle":"Placing panels in high-carbon-offset, underserved ZIPs keeps 94.6% of status-quo energy output.","key_machinery":"Carbon offset potential per panel—the estimated kilograms of CO2 avoided by replacing grid electricity with rooftop PV at a location, computed from Project Sunroof's energy-generation potential and eGRID non-baseload emission rates—is the yardstick that exposes the inefficiency. Realized potential, the ratio of existing to possible installations, is the yardstick of equity. The mechanism that carries the policy claim is the Round Robin simulation: ZIP codes are sorted by energy potential, carbon offset potential, Black population proportion, and median income, and the simulator assigns new panels by cycling through the top ZIP in each list, then the second, and so on, giving each objective equal weight. This produces a more geographically spread allocation than optimizing any single objective and is the strategy whose carbon gain and energy retention are reported.","core_discovery":"The central claim is that the current distribution of residential rooftop PV is carbon-inefficient: the US has been putting panels where they generate the most electricity, but those places often sit on grids that are already relatively clean, while places with dirtier grids—disproportionately Black, low-income, and Republican-leaning areas—have higher carbon offset potential and far fewer panels. The paper quantifies this as a mismatch between realized potential (existing installations as a share of possible installations) and carbon offset potential per panel. It then shows that a 'Round Robin' siting strategy, which gives equal weight to energy generation, carbon offset, Black population share, and low income by cycling through ZIP codes ranked on each objective, would achieve 94.6% of the status quo's energy addition while increasing carbon reductions by at least 39.8%, and would lift realized potential in high-Black-population states from about 50% below the national average to about 25% above it.","pith_inferences":["The 39.8% figure compares strategies at equal panel counts; if targeted incentives also increase total adoption in underserved ZIP codes, the total carbon savings could be larger than this bound, but that depends on demand responses the paper does not estimate.","The simulation treats each ZIP code's rooftop space as exhaustible, so a home-level version could reveal whether the highest-offset rooftops within a ZIP are also the least likely to be adopted, which would change how aggressively a policy must target individual households.","The authors note that Sunroof's carbon offset estimates ignore the duck-curve timing mismatch between solar generation and evening grid demand; redoing the analysis with hourly marginal emission rates could shrink or grow the gaps, though the broad regional inversion of high-offset areas being underserved would likely persist because grid mixes differ strongly by region.","Because the Round Robin gives equal weight to four objectives, it is only one point in a policy space; explicit weights chosen by policymakers could trade energy for carbon or equity differently, and a formal Pareto analysis would map those trade-offs."],"forward_implications":["Continuing the status quo to a fourfold panel increase would need 100% of the projected panels to reach the net-zero carbon offset level, while the Round Robin strategy reaches the same total carbon offset with only 69.0% as many panels.","A strategy that optimizes for carbon offset alone avoids 71.3% more carbon than the status quo for the same number of panels, showing that pure carbon targeting is even more aggressive on climate.","Strategies that target racial or income equity also beat the status quo on carbon, so in the simulated range the equity and climate objectives reinforce rather than oppose each other.","The energy-maximizing siting strategy performs worst on carbon despite generating the most electricity, implying that optimizing residential solar for kilowatt-hours alone can undermine decarbonization goals.","The Round Robin strategy keeps 94.6% of the status quo's added energy while spreading installations more evenly across the southeast and Midwest rather than concentrating them in the sunny southwest."],"supporting_citations":[{"why":"Provides the core dataset: per-ZIP solar energy generation potential, carbon offset potential, and existing installation counts.","marker":"[37]"},{"why":"Supplies the non-baseload emission rates used in Sunroof's carbon offset calculations, the basis for the carbon-efficiency yardstick.","marker":"[57]"},{"why":"Supplies ZIP-level race and income demographics from the American Community Survey used for the equity analysis.","marker":"[11]"},{"why":"Supplies state-level 2020 presidential election results used to classify states by Republican voting share.","marker":"[20]"},{"why":"Provides industry projections of future US residential panel counts that set the simulation's N values and net-zero scenarios.","marker":"[2]"}],"fun_headline_variants":["Round-robin solar siting: 39.8% more carbon cuts, 5.4% energy loss","Fair siting of solar panels boosts carbon efficiency 39.8%","Solar's inequity has a carbon cost; fixing it cuts CO2 39.8%","39.8% carbon savings from equitable solar siting","Siting solar where carbon matters: 39.8% more climate benefit"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The analysis assumes that Project Sunroof's carbon offset potential—which is based on yearly average eGRID non-baseload emission rates and does not correct for the mismatch between when solar generates and when the grid is dirtiest—is an accurate ranking of where rooftop panels avoid the most carbon; if that estimate is biased by region, the observed carbon inefficiency and the 39.8% improvement could both be artifacts.","fun_headline_variants_meta":{"raw":{"variants":["Round-robin solar siting: 39.8% more carbon cuts, 5.4% energy loss","Fair siting of solar panels boosts carbon efficiency 39.8%","Solar's inequity has a carbon cost; fixing it cuts CO2 39.8%","39.8% carbon savings from equitable solar siting","Siting solar where carbon matters: 39.8% more climate benefit"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001222,"raw_usage":{"total_tokens":5082,"prompt_tokens":1056,"completion_tokens":4026,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":672,"completion_tokens_details":{"reasoning_tokens":3918}},"tokens_in":672,"tokens_out":4026,"duration_ms":25354,"temperature":1.0,"reasoning_tokens":3918,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-10T15:31:00.143091+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Recompute per-ZIP carbon offset potential using hourly marginal grid emission rates—the emissions of the actual plants that would ramp down when solar generates—instead of yearly average non-baseload rates, and rerun the siting simulations; if the resulting carbon offset rankings stop showing high-offset, low-installation ZIPs, or if the Round Robin strategy no longer beats the status quo in avoided carbon, the central claim is refuted.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the core dataset: per-ZIP solar energy generation potential, carbon offset potential, and existing installation counts."},{"cited_title":"Environmental Protection Agency","cited_arxiv_id":null,"evidence_quote":"Supplies the non-baseload emission rates used in Sunroof's carbon offset calculations, the basis for the carbon-efficiency yardstick."},{"cited_title":"Census Bureau","cited_arxiv_id":null,"evidence_quote":"Supplies ZIP-level race and income demographics from the American Community Survey used for the equity analysis."},{"cited_title":"https://seia.org/research-resources/solar-market-insight-report-q2-2024/","cited_arxiv_id":null,"evidence_quote":"Provides industry projections of future US residential panel counts that set the simulation's N values and net-zero scenarios."}],"review_version":1}