new_entrants_summary v7.5/7.8: NSA pumped-hydro rows have mismatched Regional build cost zone (CSA instead of NSA) [+workbook weirdness]

[EllieKallmier]

In the v7.5 IASR workbook's new_entrants_summary table, three pumped-hydro rows with subregion = NSA have their Regional build cost zone column set to CSA instead of NSA. This has passed through to v7.8 workbook as well.

I'm pretty confident that this is a typo and not intentional:

• in the 'Regional Build Costs Summary' table PHES rows with subregion=NSA also have NSA listed as the 'Cost zone' value, and
• I haven't been able to find any indication in reports or elsewhere in the workbook that the out-of-place 'CSA' value in the summary table is intentional.

Current impact:
In the templater for new entrants, I'm treating this as a known typo rather than a real cost-zone split. The templater hardcodes the exception and keys LCF lookups on geo_id for all technologies (in this case, PHES in NSA -> NSA) regardless. PR to come shortly.

I've opened this issue partly to document this decision and partly to remind myself to follow up (see below note).

Note: I haven't yet reported this or followed up with AEMO to confirm whether this is a typo or intentional - we have previously chatted about putting together a list of workbook confusions/bugs/clarifications to put to the ISP team in one go rather than piecemeal, so I think this would go on that list. But not sure if we've started that list yet, so will also note the latest updates in the isp-workbook-parser and related documentation of changes from v7.5->7.8 that can be found here and here.

[EllieKallmier]

RELATED (workbook data requiring manual handling): In the technology_specific_lcfs table (again, both v7.5 and 7.8), most values are calculated in-place in the workbooks and returned as factors but displayed as percentages (by Excel cell formatting). So the isp-workbook-parser reads those values as factors, i.e., values between 0-1.0 (though sometimes slightly above 1.0 to indicate higher-than-baseline LCF).

However: in classic BOTN special rules, the value for the named 'BOTN - Cethana' new entrant is manually typed in the workbooks as the literal percentage value 100.00 (also displayed by Excel as % so in the workbook looks the same as the others). This means though that the workbook parser reads it in as 100.00, so the templater needs to explicitly manage this scale difference.

I'm adding this note here rather than opening a new issue to keep a few of these workbook typo/small manual handlings in the same place for the moment but might move over to a discussion/.md file somewhere to compile a more comprehensive list.

[EllieKallmier]

Another one found: 'WOO' variants of Gas new-entrants.

For all variants of the new-entrant generators with Fuel Type 'Gas' (OCGT Small, OCGT Large, CCGT, CCGT with CCS) there are two rows in the summary tables defined for subregion=CNSW.* This is not a big deal because they have different names; however, they also have almost exactly the same values throughout so it looks like a direct double.

*This is kinda explicitly mentioned in the workbook changelog for 7.4 ("Summary Mapping: [...] and change the subregions for WOO OCGT Small, WOO OCGT Large, WOO CCGT, WOO CCGT with CCS to CNSW.") and 7.5 ("Summary Mapping and all related tabs: Updated the build cost zone assignment for WOO OCGT Small, WOO OCGT Large, WOO CCGT, WOO CCGT with CCS to CNSW.").

EXCEPT:

• the MLF mappings are different: 'CNSW [name]' = NE_CNSW, 'WOO [name]' = SE_SNW
• the Connection cost - Region is different: 'CNSW [name]' = CNSW, 'WOO [name]' = SNW

These 'SNW' mappings seem to contradict the changelog notes so maybe they just weren't updated, or maybe it's intentional, but it's not clear either way. It's also worth noting that in the derived summary tables (sheet name 'New Entrant Data Summary') those mappings are never actually used:

• MLF mappings use the 'IASR ID' column directly (name), and the MLF table defines explicit 'WOO [name]' and 'CNSW [name]' mappings to existing generators - that, notably, are the same!! (Tallawarra PS).
• Connection costs map to a Technology Type and Region, and the connection cost tables for non-REZ technologies are defined at the Region level - so the subregional values would map to the same region anyway.

Also of note: nowhere in any of the ISP documents have I been able to find what 'WOO' might stand for or represent. There's nothing written up in the documents that explains why these extra 'WOO' Gas generators are new entrant options or how they're different to the other CNSW generators. So it's mostly just a mystery at this stage - as such, I'm going to keep them as they are except where leaving them both in would skew a mean over regional aggregation (i.e., when aggregating up to regional or nem-level inputs, they'll be 'deduped' before averaging so it's not a weighted average that isn't explicit/flagged clearly).

[dylanjmcconnell]

My guess is that it is Wollongong - (.. no idea why it would be abbreviated as WOO). Tallawarra is basically at Wollongong I think (perhaps why they share the MLF).

Also check out this (from an appendix) - perhaps this justified a split for GPG within CNSW? :

[EllieKallmier]

Another: some hydro power stations (and maybe units individually?) are given with 'Technology Type' == 'Hydro' (rather than 'Pumped Hydro (Xhrs storage)') when they are described in the IASR Appendix A2 as Pumped Hydro schemes. These are: Wivenhoe, Shoalhaven, and maybe/a few units of Tumut 3*.

*Tumut 3 has 6 units in total, only 3 of which are capable of some pumping. This complicates the properties of the units/power station a bit - storage capacity in hours isn't as straightforward as other closed-system PHES-only units. There are also some inconsistencies in how the capacity of the units/power system in total are reported across different tables in the workbooks:

• pumped_hydro_existing_committed_anticipated_additional_properties lists a Power Station total figure of 1800MW installed capacity -> under the name 'Lower Tumut'
• maximum_capacity_existing_committed_anticipated_additional_generators lists individual units (1-6) with equal 250MW installed capacities, for a total of 1500MW for Tumut 3 Power Station. (There are another set of 4xTumut1 and 4xTumut2 units under the Power Station name 'Upper Tumut')
  So this one is a bit messy atm...

[EllieKallmier]

Year coverage by tables doesn't always line up - noticed in particular when comparing the tables connection_cost_forecast_other / connection_cost_forecast_wind_and_solar (range: 2024/25 - 2054/55) against efficient_level_of_system_strength_cost (range: 2025/26 - 2053/54).

For the v7.5 workbook tables (post-parsing), a sweep from Claude found the following:

1. Start year differs (all ending 2054-55)

start	#	which
2023-24	3	connections_forecasts_* (all three scenarios)
2024-25	72	all 42 flow_path_augmentation_costs_*, all 15 rez_augmentation_costs_*, appliance_uptake_forecasts_*, connection_cost_forecast_*, energy_efficiency_*, distribution_network_CER_augmentation_tranche_costs_forecast, residential_electricity_price_indices
2025-26	48	everything else on the standard horizon - fuel prices, hydrogen_*, electrification_*, data_centre_forecasts_*, CER tables (ev_*, v2g_*, pvnsg_*, rooftop_pv_*), dsp_summer

2. End year differs

table(s)	span	note
gross_state_product_*, household_disposable_income_* (6)	[2025-26, 2055-56]	one FY longer than the standard horizon
build_costs	[2024-25, 2053-54]	one FY shorter - 30 columns where its 2024-25 peers have 31
efficient_level_of_system_strength_cost	[2025-26, 2053-2054]	one FY shorter and mixed notation (see 4)
long_duration_outages_existing_generators, other_outages_existing_generators	[2025-26, 2034-35]	10 years only
seasonal_ratings_existing_committed_anticipated_additional_generators	[2024-25, 2033-34]	10 years, in compound headers (see 6)
distribution_network_cer_generation_limit_profiles	[2026-27, 2049-50]	24 years, on the row axis

3. Notation changing mid-table

• efficient_level_of_system_strength_cost - runs 2025-26 … 2031-32, then switches to 2032-2033 … 2053-2054 for the remaining 22 columns. [NOTE: this is handled currently]
• synchronous_unit_commitment_requirement_accelerated_transition - _Year column runs 2025-26 … 2035-36 then ends 2036-2037. Its sibling synchronous_unit_commitment_requirement_step_change&slow_growth uses YYYY-YY throughout, so two otherwise-parallel tables disagree on the last row.

[EllieKallmier]

Another Claude-assisted review of the workbook version 7.5 tables re: geographical identifiers, focusing on REZ ID use and consistency across tables. Includes stuff that I checked manually in v7.8 tables - see my notes in italics+square brackets :)

One case where REZ Name != REZ ID

HAYMSF1 in existing_committed_anticipated_additional_generator_summary has REZ Location = Isaac but REZ ID = Q10. Isaac is Q4; Q10 is Collinsville. [Same in v7.8. On a map, this solar farm is located between Mackay and Townsville, just north-west of Proserpine; looks like Q10 from that?]

REZ ID ↔ sub-region/region disagreements with the canonical table

Checking each summary table's Region/Sub-region against what renewable_energy_zones says the
REZ's region/sub-region is (Q8a/b/c normalised to Q8 first):

existing_committed_anticipated_additional_generator_summary — 6 rows

generator	REZ ID	table says	canonical says
CSPVPS1	Q10	QLD/CQ	QLD/NQ
DAYDSF1	Q10	QLD/CQ	QLD/NQ
EMERASF1	Q4	QLD/NQ	QLD/CQ
WPWF	S4	SA/NSA	SA/CSA
Eraring Big Battery	N9	NSW/SNW	NSW/CNSW
Bullyard Solar Farm	Q7	QLD/CQ	QLD/SQ

[The three in Q10/Q4 here could be same kind of thing as above - Q10 and Q4 sit geographically side-by-side, with Q10=NQ and Q4=CQ, so looks like either there's a generator-specific assignment happening here or a small typo mixing up the two].

new_entrants_summary — 7 rows, all N13

Every N13 row (N13_SAT/CST/WH/WM_South Cobar and the three N13 Battery rows) says NSW/CNSW
where renewable_energy_zones says N13 is in NSW/SNSW. Systematic across all 7, so this reads like a
deliberate assignment rather than a typo — but it does mean new_entrants_summary's Sub-region
and its REZ ID disagree for N13.

[In the IASR 2025: Table 35, N13 is listed int he CNSW ISP sub-region with reference node Wellington. Based on this and a vague comparison between network representation maps - looks to me like N13 is intended to sit in CNSW subregion.]