Cross-border river assessments diverge despite European harmonization

A comparison of ecological monitoring carried out independently by Portugal and Spain found that official assessments of shared rivers did not always align, despite both countries operating under the European Union Water Framework Directive. Across 32 water bodies in the Minho, Lima, Douro, Tagus and Guadiana basins, agreement depended strongly on which biological group was used: fish produced the closest alignment between national systems, while diatoms showed the sharpest divergence. The study analyzed data collected from 2019 to 2021 and indicates that formal European intercalibration of classification boundaries did not by itself ensure operational comparability where monitoring procedures, taxonomic practices and national reference conditions differed.

The analysis covered 33 sites, including 21 boundary water bodies and 12 transboundary water bodies distributed among 16 sub-basins. Seven sites were sampled twice, producing 40 monitoring records. The datasets included 23 fish records per country, 18 macrophyte records per country, and 37 records per country for both benthic macroinvertebrates and diatoms. Eight sampling records contained all four biological quality elements, 21 contained three, nine contained two and two contained only one. The sampled rivers ranged from 2.7 to 80.92 km in length and had drainage areas of 7.93 to 259.22 square kilometres, so all were small- to medium-sized, wadeable river systems rather than large rivers.

National teams collected and evaluated their data separately under their respective Water Framework Directive programmes. Although both systems used the same broad regulatory framework and had undergone intercalibration, the protocols differed in sampling season, spatial design, field procedures, laboratory processing and the indices used to translate biological observations into ecological status classes. Portugal and Spain also maintained separate type-specific reference conditions against which ecological quality ratios, or EQRs, were calculated. These ratios compare observed biological values with those reference conditions, while the final status of each site was determined by the “one-out, all-out” principle: the lowest class assigned to any assessed biological quality element became the site’s overall classification.

The researchers used three complementary methods to test where national assessments agreed or diverged. Nonmetric multidimensional scaling and an analysis of similarities examined the composition of biological assemblages. EQR distributions were compared with paired Wilcoxon signed-rank tests, and Kendall’s tau rank correlations assessed agreement among national ecological classifications. Fish assemblages were analyzed at species level, macrophytes at species level, macroinvertebrates at family level and diatoms at genus level. Harmonizing the diatom records at genus level was necessary because the Portuguese and Spanish teams had used different taxonomic resolutions, although the researchers cautioned that this reduced the biological detail available in the comparison.

Fish communities showed extensive overlap between Portugal and Spain. The analysis of similarities detected no statistically significant compositional difference between the countries, with an R value of −0.005 and p = 0.466. Their EQR values tended to be lower in Spain than in Portugal, but that paired difference did not reach conventional statistical significance, at p = 0.057. The ecological classes assigned to the same fish assemblages had a moderate positive correlation across borders, with Kendall’s tau of 0.40, the only statistically significant positive between-country correlation among the four biological groups. Even so, the two systems assigned different fish classes to 86.9% of the compared water bodies, showing that a significant correlation did not mean classifications were routinely identical.

Macrophytes and macroinvertebrates occupied an intermediate position. Portuguese and Spanish macrophyte assemblages showed limited overlap in the ordination, and the analysis detected a statistically significant difference, with R = 0.409 and p = 0.001. The ordination had a relatively high stress value of 0.24, however, indicating that its two-dimensional representation should be interpreted cautiously. Their EQR distributions were nearly identical, with no significant difference between countries at p = 1.000, and the corresponding classifications had a weak, nonsignificant positive correlation of tau = 0.20. Macroinvertebrate assemblages overlapped partially and showed weak but statistically significant separation, with R = 0.109 and p = 0.002. Spanish EQR values tended to be higher, although the difference was not significant at p = 0.055; the cross-country classification correlation was weak and nonsignificant at tau = 0.30. Pairwise disagreement was 44.4% for macrophytes and 59.5% for macroinvertebrates.

Diatoms were the most strongly divided element. Portuguese and Spanish diatom assemblages formed distinct clusters with minimal overlap, and the analysis of similarities produced the largest separation among the four groups, with R = 0.635 and p = 0.001. This was the only biological group whose paired EQR values differed significantly between countries: Spanish values were higher, with p below 0.001. Yet the ecological classes derived from those EQRs had a very weak, nonsignificant negative correlation across the border, at tau = −0.10, and the systems disagreed in 70.3% of paired diatom assessments. The combination points to differences at both the biological-record and assessment stages rather than to a single uniform shift in measured condition.

The biological indicators also influenced one another differently within the two national systems. In Portugal, fish and diatom classifications were significantly negatively correlated, at tau = −0.60, suggesting that the two indicators ranked sites differently. Portugal’s final classification was positively associated with its fish and macroinvertebrate classifications, with tau values of 0.50 and 0.30. In Spain, fish and macroinvertebrate classifications had a significant positive correlation of tau = 0.50. More strikingly, Spain’s fish classification was perfectly correlated with its final ecological classification, at tau = 1.00, while its final class was also significantly associated with macroinvertebrate and diatom classifications, both at tau = 0.40. The “one-out, all-out” rule could therefore amplify a divergence already present in a single biological element.

When all available indicators were combined, Portugal and Spain produced a final ecological classification with only a weak, nonsignificant positive correlation, at tau = 0.30, and disagreed on 80.0% of the overall assessments. The authors said this result demonstrates that formal intercalibration, which harmonizes class boundaries and seeks comparable levels of ecological alteration, cannot guarantee equivalent outcomes at the scale of shared water bodies. Their findings instead identify sampling timing, spatial coverage, sample treatment, taxonomic resolution, index construction and national reference conditions as possible sources of operational divergence. The study did not quantify how much disagreement arose from biological differences versus assessment procedures, and it did not include directly comparable physicochemical or hydromorphological indicators.

For transboundary management, the central implication is that a shared legal framework can conceal marked differences in how ecological condition is measured. Divergent classifications may complicate priority setting, restoration planning and compliance reporting for the same connected rivers. The authors therefore called for closer harmonization of monitoring procedures, taxonomic practices and assessment protocols between neighboring countries, especially for the biological elements that showed the greatest divergence. Their results also caution that agreement must be tested at individual sampling sites and for each biological quality element, rather than inferred solely from successful European intercalibration. At the same time, the analysis covered a limited set of small- and medium-sized rivers, excluded highly modified reservoirs and used historical monitoring records, so its findings should not be treated as a direct measure of ecological improvement or decline. It is instead a detailed demonstration that harmonized environmental reporting can still produce inconsistent river classifications when the underlying field and laboratory practices are not aligned.

Subject of Research: Operational comparability of biological ecological-status assessments in Portuguese and Spanish transboundary rivers

Article Title: Operational comparability of ecological status assessment in Iberian transboundary rivers

Article References: Tovar Faro, A., Oliveira, J. M., Duarte, G., Branco, P., Segurado, P., & Ferreira, T. (2026). Operational comparability of ecological status assessment in Iberian transboundary rivers. Environmental Monitoring and Assessment, 198(11), Article 1226. https://doi.org/10.1007/s10661-026-16021-6

Image Credits: AI Generated

DOI: 10.1007/s10661-026-16021-6

Keywords: Water Framework Directive, Iberian rivers, Ecological status, Biological monitoring, Portugal and Spain, Diatom assemblages, Fish communities, Intercalibration, Transboundary water management, Operational, comparability, ecological