In brief

  • The Ministry of Development has approved a single methodology for drawing up seismic risk maps, and from 11 October 2026 every new map has to follow it. The result is no longer a coloured area on a county, but a value calculated for each census sector, that is for areas of a few hundred dwellings, expressed as a percentage of the replacement value of the buildings there.
  • The methodology comes with a table of 3,181 rows, one for each administrative-territorial unit, in which every locality receives its design ground acceleration. 1,080 localities, 34% of the total, have a value of at least 0.30g, and 126 of them reach the maximum in the table, 0.50g. Bucharest appears with 0.37g.
  • The calculation goes all the way to money. For a four-storey block of flats built in 1986 in Iași, with 1,400 square metres of gross floor area, the worked example in the act gives expected losses of 318,249 euros for the earthquake with a return period of 225 years and of 760,324 euros for the one with 975 years, out of a replacement cost of 2.24 million.
Act: Order of the Minister of Development, Public Works and Administration no. 1.037 of 3 September 2026 approving the technical regulation „Methodology for drawing up seismic risk maps, indicative RTC 18-2026”
Published: Official Gazette of Romania (Monitorul Oficial) no. 771 of 11 September 2026
In force from: 11 October 2026, 30 days after publication

From 11 October, Romania has an official recipe for calculating the damage an earthquake would cause, applicable at neighbourhood level. Order no. 1.037/2026 of the Minister of Development, published in Official Gazette of Romania no. 771 of 11 September 2026, approves the technical regulation RTC 18-2026, and the methodology itself takes up 149 pages in the bis issue of the same edition. It joins a field that has just been rewritten from the ground up: since 25 August the Spatial Planning, Urban Planning and Construction Code has applied, repealing Law no. 50/1991 and Law no. 350/2001.

Government Decision no. 447/2003, which the methodology itself lists among its reference documents, approves the content rules for landslide risk maps and for flood hazard and risk maps. For earthquakes there was money, through Government Decision no. 932/2007, which covers up to half the cost of the map from the State budget, while the same list of reference documents contains no earlier technical regulation on the content of seismic risk maps. That is where the differences came from: two town halls could commission two studies that measured different things and looked different.

The methodology is built on four layers that multiply one another. The first is the hazard, that is how strongly the ground moves. The act reduces it to a single quantity, the peak horizontal ground acceleration, written PGA, and to three scenario earthquakes: one with a probability of exceedance of 20% in 50 years, which returns on average once in 225 years, one of 10% in 50 years, at 475 years, and one of 5% in 50 years, at 975 years. The starting values are not to be looked for, they are to be read off: Table A.1 in the annex gives, for each of the 3,181 administrative-territorial units, the acceleration corresponding to the middle scenario and the slope of the hazard curve, while the other two scenarios follow from them through two formulas written into the act.

The second layer is exposure, that is the inventory of buildings. This is where the methodology demands the most fieldwork. Every building in a locality, whatever its use, is placed in a typology made up of four elements: the year of construction, split into seven classes, from those put up before 1940 to those after 2013; the structural system, split into twelve types, from unreinforced masonry with timber floors to dual steel structures; the height class, in three bands; and the use, in four groups. The data are gathered from field inspections, from the land registry records, from the rapid visual assessment carried out under regulation RTC 10-2022, from questionnaires sent to owners and, the act says, from satellite images, drones and 3D scanning.

The third layer is fragility. For each structural typology, Annex B gives the acceleration thresholds at which the building reaches one of the four damage states the methodology uses: slight, moderate, extensive and complete. From these, using a lognormal distribution function, the probability that a building is in each state under the chosen scenario earthquake is calculated. The fourth layer turns the probabilities into money, through a table of fixed losses: slight damage costs 5% of the replacement value of the building, moderate damage 15%, extensive damage 60% and complete damage 100%.

The replacement value is not estimated case by case but taken from a second table, per square metre of gross floor area and per use, with prices referenced to 2025 and updatable with the construction cost indices published by the National Institute of Statistics. A single-family home is counted at 1,400 euros per square metre, a block of flats at 1,600, an administrative building at 2,000, a retail building at 2,200, and a hospital at 3,500, the most expensive on the list.

The worked example in Annex C shows how concrete the result becomes. A block of flats in Iași, built in 1986, with a reinforced concrete frame structure, basement, ground floor and four upper storeys, with 1,400 square metres of gross floor area, has a replacement cost of 2,240,000 euros. For the earthquake with a 225-year return period, the expected loss is 14.21% of that value, that is 318,249 euros. For the one with a 475-year return period it rises to 22.83%, and for the one with a 975-year return period it reaches 33.94%, that is 760,324 euros.

What it changes in practice

The first effect is on what town halls buy. The methodology is a technical regulation in construction and spatial planning, so its requirements become mandatory for any seismic risk map drawn up or updated after 11 October 2026. Annex D adds a model set of tender specifications, with CPV code 71351914-3, in which the beneficiary is the administrative-territorial unit and the contractor has to bring a team covering three specialities: civil engineering, urban planning and geodetic engineering.

The second effect is on the form of the result. The map of a commune or a town is drawn at scales between 1:1,000 and 1:5,000 and shows, for each census sector, two things: the expected damage state and the expected percentage of loss. The colours are fixed in the act, hexadecimal codes included, from deep green for losses under 5% to deep red for over 60%. The map of a county is drawn at scales between 1:25,000 and 1:100,000 and shows something else: the risk indicator and the resilience indicator, calculated for each locality in the county. The national map runs at scales between 1:250,000 and 1:1,000,000 and aggregates the counties.

The third effect concerns two new indicators, which make rankings possible. The seismic risk indicator of a locality is the ratio between the expected losses and the total replacement cost of the buildings there, hence a percentage. The resilience indicator is the ratio between the same losses and the gross domestic product of the locality, so it shows how much of the local economy would be needed to repair the damage. Localities and counties are ranked, according to the act, by the values calculated for the middle scenario, the one with a 475-year return period.

The fourth effect is technical, but it decides whether the maps can be laid on top of one another. The deliverables are GeoTIFF files, shapefiles and geodatabases, plus a PDF report, in the ETRS89/LAEA Europe, WGS84 or Stereo 70 reference systems. The contracting authority has to ensure interoperability with the INSPIRE specifications for natural risk zones and with the technical rules on spatial data sets in urban planning documentation.

The fifth effect is the one that is missing. The methodology says how the map is made, not when. No provision of the order or of the annex sets a date by which a locality must have its seismic risk map, and drawing one up remains a decision of the local authority, financed from its own budget, from European funds or, up to 50%, from the State budget through the programme in Government Decision no. 932/2007.

What has changed compared with the previous situation

The most visible change is the move from zoning to calculation. The national spatial plan approved by Law no. 575/2001 works with geographically delimited areas, inside which there is a potential for destructive natural phenomena. The new methodology does not delimit areas, it calculates a figure: the probability of damage, then the loss in euros, then the percentage of the value of the building stock. The risk zone remains the legal instrument, but its content becomes a numerical result, not an assessment.

The second change is that the age of the building enters the calculation with an explicit weight. The thresholds in Annex B are multiplied by a factor that depends on the year of construction and on the locality. Buildings put up after 2013 receive a factor of 1.5 times the local scaling factor, those from 2007 to 2013 a factor of 1.2 times, and those from 1978 to 2006 the local factor itself. For buildings from 1964 to 1977 the factor no longer depends on the locality but on the county, with three values: 0.67 in nineteen counties, 1.00 in fifteen and in Bucharest, 1.50 in Bacău, Brăila, Buzău, Galați, Prahova, Vaslui and Vrancea. For buildings put up between 1940 and 1963, the factor also depends on the use they had at the time: public ones or private ones of public interest, such as theatres, hospitals, schools and hotels, receive 1.50 in one group of counties and 2.25 in Bacău, Brăila, Buzău, Galați, Prahova, Vaslui and Vrancea, while all the others receive 1.00.

The third change is where the result ends up. Maps that have been endorsed and approved are included in spatial planning and urban planning documentation, in order to underpin three distinct things: the measures for preventing seismic risk, the rules for putting up buildings and using land, and the defence and intervention plans in the event of an earthquake. The link with civil protection is written the other way round too: the maps can feed into the county risk analysis and coverage plans.

The fourth change is one of context, and the act does not record it. The list of normative reference documents in chapter 1.8 includes Law no. 50/1991 on authorising the execution of construction works and Law no. 350/2001 on spatial planning and urban planning, both repealed in full on 25 August 2026, 17 days before the order was published. The explanation can be read in the preamble itself: the two endorsements that closed the text date from 26 January and 22 May 2026, and the Code appeared in the Official Gazette of Romania only on 10 August.

Advantages and disadvantages

What it improves

  • It removes the arbitrariness from the commission. Two town halls that now buy studies from different firms receive comparable results, because the input data, the damage thresholds and the prices per square metre are fixed in the annexes, not left to the designer’s choice.
  • It gives every locality a ready-calculated starting value. Table A.1 spares the local authority a seismic hazard study of its own, which is the most expensive and the most contestable part of such an analysis.
  • It brings the resolution down to the census sector, so the map shows differences inside the same town, between an interwar neighbourhood and one of new blocks, instead of a single colour over the whole locality.
  • It expresses risk in money and in percentages, not in labels. A local council comparing an expected loss of 15% with one of 45% has something to prioritise its strengthening works by, and the comparison with local gross domestic product also shows how hard the locality would find it to recover afterwards.
  • It fixes the format of the deliverables and the coordinate systems, which means the maps can be overlaid on other layers and aggregated at county and national level, instead of each remaining a separate PDF.
  • It acknowledges the limits of the method outright. The text provides that the three scenario earthquakes are conventional, that they are not predictions and that the results are to be compared between localities, not with the effects observed after a real earthquake.

What remains a problem

  • Annex B, which gives the damage thresholds, contains no column for buildings put up between 2007 and 2013 and has a single column for those after 2013, for one structural system out of twelve. For the rest of the building stock of the past twenty years, the value that has to be multiplied by the factor in the act does not exist.
  • The six sectors of Bucharest are declared level IV units, so they have to have maps of their own, yet they are expressly excluded from the table of accelerations. The capital appears there with a single value, 0.37g, given as a county-level value.
  • There is no deadline at all for drawing up the maps. The methodology gives itself 30 days until it enters into force, but says nothing about when a locality in the 0.50g band has to have a map.
  • Replacement costs are fixed for 2025 and are to be updated with the construction cost indices. The act does not say who does the updating, nor what happens to maps drawn up with old prices, even though the colour on the map hangs directly on that value.
  • The inventory required is enormous. Every building has to be classified by year of construction, structural system, height class and use, and for many localities these data do not exist anywhere, so the real cost of the map is the cost of the inventory, not of the calculation.
  • The result remains an estimate with moderate uncertainty, acknowledged as such in the act, yet it ends up in urban planning documentation, where it is read as a technical fact. The distance between those two readings is not handled by any provision.

Practical advice

  1. If you are a town hall with a procurement procedure already under way, check the date on which you sent the contract notice for publication or issued the invitation. Under Article 2(b), contracts and documentation for which the procedure was launched before the date of entry into force stay outside the new requirements, therefore on the rules in the existing tender specifications.
  2. Before commissioning the study, read the value for your locality in Table A.1 of the annex. The acceleration and the slope of the hazard curve are fixed input data, and a bidder proposing a hazard study of its own is selling you work the methodology no longer asks for.
  3. Prepare the inventory of buildings before the tender, not after. The act allows the use of data from the land registry records, from the census of the building stock, from the ministry’s rapid visual assessment platform and from the authority’s own databases built under Law no. 212/2022. The more data you bring in already collected, the smaller the expensive part of the contract.
  4. Ask for the deliverables in the formats set out in the act, not just the report. Annex D lists vector files, geodatabases and geospatial files compatible with the National Territorial Observatory and with the INSPIRE geoportal. A study delivered only as a PDF cannot be aggregated at county level and will have to be redone.
  5. If you own or manage a block of flats, note that the map is not a technical survey of your building. It works with typologies and with averages per census sector, and placing a particular building in a seismic risk class is still done through a technical survey, under Code P 100-3/2019.
  6. If you are drafting tender specifications, do not take the list of reference documents from Annex D as it stands. Two of the acts there, Law no. 50/1991 and Law no. 350/2001, have been repealed since 25 August 2026, and the correct references are now to the Spatial Planning, Urban Planning and Construction Code.

Frequently asked questions

What, in short, is a seismic risk map under this methodology?
It is the graphical and analytical representation of seismic risk at the level of an administrative-territorial unit, bringing together the data on ground motion, on the existing buildings and on their fragility. For a commune or a town, the map shows, for each census sector, the expected damage state and the percentage of loss from the replacement value of the buildings.
From when does it apply?
From 11 October 2026, that is 30 days after publication in Official Gazette of Romania no. 771 of 11 September 2026, under Article 3 of the order. The order does not apply to certain older contracts, however: financing contracts concluded before that date under the State budget programme, service contracts for drawing up or updating the maps, and contracts or documentation for which the procurement procedure had already been launched.
What are the three scenario earthquakes?
They are three conventional levels of severity, defined by the probability of being exceeded in 50 years: 20%, 10% and 5%. The corresponding average return periods are 225, 475 and 975 years. The act says expressly that these are calculation conventions, not predictions of real events, and that they cannot be compared with the way an earthquake behaves at a particular site.
What do the four damage states mean?
They are degrees of deterioration of a building after an earthquake: slight, moderate, extensive and complete. The methodology attaches fixed losses to them, expressed as a percentage of the replacement cost: 5% for slight damage, 15% for moderate, 60% for extensive and 100% for complete. The probability of each state is calculated from the acceleration thresholds given in Annex B.
Who pays for the map?
The local authority, as contracting authority. It can be financed from the local budget, from European funds or, through the programme governed by Government Decision no. 932/2007, up to 50% from the State budget, through the budget of the Ministry of Development. The methodology sets no deadline by which the map has to be produced.
Does the map tell me whether my block of flats is at seismic risk?
No. The analysis works on structural typologies and gives weighted averages per census sector, that is over areas containing many buildings. Placing a particular building in a seismic risk class remains the result of the technical survey carried out under Code P 100-3/2019, and the act points out that the results of the risk analysis are relevant mainly for comparisons between localities.
Why does the year of construction matter so much?
Because it shows which design rules the building was conceived under. The methodology uses seven classes, from those before 1940 to those after 2013, and multiplies the damage thresholds by a factor that rises for newer buildings. A building from after 2013 receives a factor of 1.5 times the local scaling factor, against the plain factor for those from 1978 to 2006.
Which locality has the highest value in the table?
Not just one. The maximum value in Table A.1, 0.50g, appears for 126 localities, all in four counties: 59 in Buzău, 55 in Vrancea, 10 in Galați and 2 in Brăila. At the other end, 430 localities have the minimum value, 0.10g. Bucharest is listed with 0.37g, with the slope of the hazard curve equal to 3.
Where can I find the full text of the methodology?
In the Official Gazette of Romania, Part I, no. 771 bis of 11 September 2026, pages 3 to 151. The order itself is published in no. 771 of the same day, on page 7, and is also reproduced in the bis issue, on page 2. The official edition is attached to this article, in PDF format.

Errors and inconsistencies in the published text

  • Annex B has no parameters for buildings constructed after 2006, although chapter 4 gives them multiplying factors. Point 3.2 (2)(f) and (g) creates class a6, for buildings put up between 2007 and 2013, and class a7, for those after 2013. Chapter 4, paragraph (4)(a) and (b) provides that their thresholds are obtained by multiplying the values in Annex B by 1.5·FHS and by 1.2·FHS respectively. In Table B.1, class a6 does not appear at all, and class a7 appears only for structural system DO, dual steel structures. For all the other eleven structural systems the value that ought to be multiplied is missing, so for buildings put up in the past twenty years the fragility cannot be calculated, and without it neither the loss nor the indicator of the locality can be obtained. The same break appears at the other end of the range: point 3.2 (3)(c) expressly includes in the ZC/ZA typology multi-storey buildings with concrete columns and beams and masonry walls built before 1940, while in Table B.1 the same typology begins only at class a3, that is at 1964.
  • Article 2(a) excludes service contracts for drawing up the maps from the scope of the order, with no limit in time. The published text says that the order does not apply to „contractele de finanțare încheiate înainte de data intrării în vigoare a prezentului ordin, în cadrul Programului privind finanțarea elaborării și/sau actualizării de la bugetul de stat a hărților de risc natural pentru cutremure și alunecări de teren, și contractele de servicii cu operatorii economici pentru elaborarea și/sau actualizarea hărților de risc natural pentru cutremure”, that is financing contracts concluded before the date of entry into force of the order, under the programme for financing from the State budget the drawing up or updating of natural risk maps for earthquakes and landslides, and service contracts with economic operators for drawing up or updating natural risk maps for earthquakes. The condition of time is written once, next to the financing contracts. The next category, the service contracts, is left without a condition of time, even though point (b) that follows sets one expressly. A reader in good faith can reach two conclusions: either the exclusion is transitional and covers only contracts concluded by 11 October 2026, or it is permanent, in which case the methodology does not apply to precisely those contracts through which the maps are drawn up.
  • The sectors of Bucharest are required to have maps, but have no hazard values. Point 5.4 (1) expressly includes them among the level IV units, alongside communes, towns and municipalities, and chapter 6.2 requires for each level IV unit a map at scales between 1:1,000 and 1:5,000. Annex A, paragraph (1), however, says that the values of the acceleration and of the slope of the hazard curve are given „pentru fiecare UAT de nivel IV, cu excepția sectoarelor municipiului București, și pentru municipiul București, ca UAT de nivel III”, that is for each level IV administrative-territorial unit except the sectors of Bucharest, and for the municipality of Bucharest as a level III unit, while in Table A.1 the capital appears once, at position 695, with 0.37g. Chapter 2, paragraph (8) refers for input data back to Table A.1. For the six sectors there is neither a value of their own in the table nor any provision saying that the value of the municipality is to be used.
  • Table 5.2 does not cover all the uses defined at point 3.2 (5). The classification of uses includes „clădiri pentru ocrotirea sănătății și asistență socială”, „clădiri pentru activități culturale, educație, cercetare, culte, sport, turism, agrement, loisir” and „clădiri cu funcțiuni agro-zootehnice”, that is buildings for health care and social assistance, buildings for cultural activities, education, research, worship, sport, tourism and leisure, and buildings with agricultural and livestock uses. The table of replacement costs per square metre, the only source for the value used in formula (5.11), instead has the rows „clădiri pentru ocrotirea sănătății”, „clădiri pentru activități culturale, educație, culte, sport, turism, agrement” and „clădiri cu funcțiuni agricole”, that is buildings for health care, buildings for cultural activities, education, worship, sport, tourism and leisure, and buildings with agricultural uses. For a social assistance building, for a research building and for a livestock building there is no price per square metre, so the replacement cost cannot be calculated, and from it come both the loss in euros and the risk indicator of the locality.

Editorial analysis

The problem the act solves is real and long-standing. Natural risk maps for earthquakes have been financed from the State budget since 2007, without a technical regulation on their content, which left it to each designer to decide what is measured and how it is shown. RTC 18-2026 closes the gap in the most useful way possible: it does not ask local authorities to produce science, it gives them the hazard data ready-calculated for all 3,181 localities and leaves them only the part that only they can do, the inventory of buildings. The decision to express risk as a percentage of the replacement value, rather than in labels, is the second good decision: a percentage can be compared, a label can be negotiated.

The figure that says the most about the act is written nowhere in it but comes out of counting Table A.1. Of the 3,181 localities, 1,080, that is 34%, have a design acceleration of at least 0.30g, 530 are at 0.40g or above, and 126 reach the maximum of the table, 0.50g. At the other end, 430 localities sit at the minimum of 0.10g. In other words, a third of the country enters the calculation with high values, and it is there that the difference between having a map and not having one costs the most. It is precisely for that third that the methodology provides no deadline and no prioritisation: the map remains optional in Focșani just as it does in Abrud.

The second observation comes from combining three places in the act and does not show up if the chapters are read one by one. The thresholds in Annex B are multiplied by the local factor FHS for buildings from 1978 to 2006, but by a flat 1.00 for those before 1940. In 555 localities the local factor is 0.33, and in another 563 it is 0.50. For a reinforced concrete frame block of five to eight storeys in a locality with a factor of 0.33, as Cluj-Napoca is, the threshold for complete damage becomes 0.736 × 0.33, that is 0.243g, while for a building in the same category put up before 1940 it stays at 0.428g. The model therefore treats the interwar building as almost twice as resistant as the one from 1985, in a third of the country’s localities. The result follows logically from the way the factor is built, since it reflects how demanding the local design code was, but not a line in the act explains it, and on the map it shows as a colour, not as an assumption.

The third observation concerns the clock, not the content. The two endorsements that closed the text date from 26 January and 22 May 2026, the order was signed on 3 September, and between those moments, on 10 August, the Spatial Planning, Urban Planning and Construction Code appeared, repealing from 25 August both Law no. 50/1991 and Law no. 350/2001. The technical regulation cites both of them in its list of reference documents. The only reason the legal basis of the order held is that Article 583(3)(c) of the Code repealed from Law no. 10/1995 Articles 1 to 9, 11 to 40 and 42 to 44, that is it left standing precisely Article 10, the one on which the issuing of technical regulations rests. That piece of luck is no substitute for an update: an order published on 11 September sends the reader to two laws that no longer existed on 25 August.

The last observation is small, but it shows how little the text was reread. In chapter 1.6, paragraph (6), which imposes an interoperability obligation on the contracting authority, the reference act is identified as Order of the Minister of Development no. 940/2023. In chapter 1.8(v) and in Annex D, the same act, with an identical title, appears as Order no. 904/2023, and in Annex D it is also given with the issue in which it was published, Official Gazette of Romania no. 470 of 29 May 2023. The transposed digits do not change the outcome, since the full title identifies the act, but the obligation is read from the paragraph carrying the wrong number.

What should be changed

  • Completing Table B.1 with classes a6 and a7 for all twelve structural systems. Effect: buildings put up after 2006, the only segment of the building stock that is growing, can be included in the analysis. Until then, any map drawn up correctly under the methodology leaves precisely the new buildings out of the calculation.
  • Repeating the condition of time in Article 2(a), for each of the two categories of contract. Effect: the reading under which the methodology would never apply to the service contracts through which the maps are drawn up, that is to its own subject matter, disappears.
  • Adding the six sectors of Bucharest to Table A.1. Effect: the capital can receive the level IV maps the act requires, and the level III indicator for Bucharest can be obtained by aggregation, as chapter 5.5 provides, instead of remaining the only value available.
  • Bringing Table 5.2 onto the same list of uses as point 3.2 (5). Effect: social assistance buildings, research buildings and livestock buildings receive a replacement cost, so they enter the total loss of the locality instead of being classified by analogy by each designer in their own way.
  • A deadline, at least for localities with a design acceleration of at least 0.40g. Effect: the 530 localities in that band would have maps within a foreseeable horizon, and the State budget financing programme could be sized against a calendar rather than against applications arriving at random.
  • A rule for updating the costs in Table 5.2, with an issuing authority and a frequency. Effect: two maps drawn up two years apart become comparable, and the colour of a census sector no longer depends on the year in which the town hall signed the contract.

Original text of the legal act

The text below is reproduced in Romanian, the official form of publication.

The full text, as published in the Official Gazette of Romania

Official Gazette of Romania no. 771 of 11 September 2026 152 pages PDF, 928 KB the act starts on page 2

Open the official PDFDownload the PDF

The viewer is not shown on small screens. Use the buttons above to open or download the file.

This article is for informational purposes only and does not constitute legal advice. For specific situations, consult a licensed attorney or tax advisor.