Ecological validation and robust applicability rules for the MTMMHC-III modified Tennant environmental-flow criteria

The Modified Tennant Method based on Multilevel Habitat Conditions Type III (MTMMHC-III; Li and Kang, 2014) derives monthly ecological flow criteria for a river purely from flow statistics: for each calendar month and wet/normal/dry water-year class, the minimum ecological flow is the daily flow exceeded 90% of the time ($Q_{90}$), the optimum range is capped at the median flow ($Q_{50}$), and intermediate classes follow an arithmetic progression whose class count is set by a mode-based rule with a tolerance of 10% of $Q_{50}$. The criteria have no demonstrated link to specific ecological responses, several parameter choices have been criticized in the later literature as subjective, and the boundary conditions for applying or transferring the criteria (required flow-record length, transfer of criteria from nearby rivers) are informal. No published study has biologically validated these criteria or established robust, evidence-based applicability or transfer rules. The open problem is to ground the criteria in ecological evidence and to replace the informal boundary conditions with validated transfer guidelines.

Background

An environmental flow (EF) is the quantity, timing, and quality of water flow required to sustain freshwater ecosystems and the human livelihoods that depend on them. Where time and data for detailed ecological study are lacking, managers rely on look-up hydrological methods, of which the Tennant method (Tennant 1976) is the most widely used: it prescribes fixed percentages of the mean annual flow — for example 10% as a short-term survival minimum and 60–100% as an optimum range — as habitat-condition classes for fish and other aquatic organisms. Its weaknesses are equally well known: the percentages were derived from a small set of United States rivers, they ignore seasonal and year-to-year flow variability, and they take no account of the specific river to which they are applied.

Li and Kang (2014) proposed one of the most elaborate modifications, the Modified Tennant Method based on Multilevel Habitat Conditions Type III (MTMMHC-III). A flow duration curve (FDC) plots the fraction of time a given daily streamflow is exceeded; $Q_{50}$ and $Q_{90}$ denote the daily flows exceeded 50% and 90% of the time. For each calendar month $i$ and each water-year class $j$ — wet, normal, or dry, defined by whether the guaranteed rate (the exceedance probability of the monthly mean flow) falls below 25%, between 25% and 75%, or above 75% — the method sets the minimum ecological flow (MEF) to $Q_{90}(i,j)$, caps the optimum range at the median $Q_{50}(i,j)$, and fills the interval with an arithmetic progression

$$E_{n}(i,j)=Q_{90}(i,j)+(n-1)\,\frac{Q_{50}(i,j)-Q_{90}(i,j)}{9},\quad n=1,\dots,10,$$

mirroring the Tennant classes from degraded to optimum. The number of classes actually used is reduced by rounding the ratio of the minimum-to-optimum span to a tolerance of $0.1\,Q_{50}(i,j)$ and taking the mode across months, which yielded 4 classes for the Hanjiang and Meijiang rivers and 3 for the Panlonghe River in the paper's case studies. Every input is a hydrological statistic obtainable from a gauged daily flow record.

The limitations that define the open problem follow from the method's construction. The criteria are built from statistics chosen because they are easy to obtain, not from biological evidence: no ecological or physical requirement enters the choice of $Q_{50}$ as the optimum ceiling, $Q_{90}$ as the minimum, the 25%/75% wet–dry thresholds, the $0.1\,Q_{50}$ tolerance, or the mode-based class count, and the classification levels have never been compared against measured ecological responses. The applicability boundary conditions — how long a flow record must be to capture full inter-annual variability or monthly variability, and when criteria may be transferred from a nearby gauged river — are informal guidance that can be modified case by case. The later literature confirms that these limitations persist: Fu et al. (2020) judge the method "only suitable to humid regions" with "wet, normal, and dry years ... subjectively determined". Habitat-simulation methods (instream-flow models that link river hydraulics to the habitat suitability of target species) could in principle supply the missing biological link, but they are rarely applied because of their data demands (Ma et al. 2019).

Problem Statement

The problem raised by the MTMMHC-III method of Li and Kang (2014) and left open by the subsequent literature is unified by one objective: making Tennant-type ecological flow criteria ecologically defensible and robustly applicable. It has two linked parts, both of which are mandatory.

First, ecological grounding of the criteria. The criteria levels $E_{1},\dots,E_{10}$ built from $Q_{90}$, $Q_{50}$, and the arithmetic progression above, together with the method's parameter choices — the guaranteed-rate thresholds 25% and 75% separating wet, normal, and dry water-year groups, the tolerance $0.1\,Q_{50}$ in the class-number rule, and the mode-based single class count per water-year group — must be linked to specific ecological responses: do rivers held at or above a given criterion level exhibit measurably better ecological condition (habitat suitability, fish or macroinvertebrate status, riparian vegetation), and on that biological and physical basis, which parameter values are justified, or how should the criteria be revised — including cases where months within one water-year group require different class counts, which the single mode-based number cannot capture?

Second, robust applicability and transfer rules. The boundary conditions under which the criteria may be applied or transferred must be made robust rather than informal: what flow-record length is actually required, for which river types and flow regimes do the criteria hold, and under what conditions can criteria be transferred to a river without an adequate record from a nearby gauged river? For the first part, the ecological grounding must be argued from existing evidence or from a validation design — the ecological comparisons to be made and the justification procedure for each parameter — rather than from a completed empirical validation. For the second part, the applicability and transfer rules must be stated as general conditions rather than only for the Hanjiang, Meijiang, and Panlonghe case rivers.

References

1. Li, Changwen; Kang, Ling (2014). 'A New Modified Tennant Method with Spatial-Temporal Variability'. Water Resources Management 28(14): 4911–4926. doi:10.1007/s11269-014-0746-4. https://doi.org/10.1007/s11269-014-0746-4
2. Tennant, Donald Leroy (1976). 'Instream Flow Regimens for Fish, Wildlife, Recreation and Related Environmental Resources'. Fisheries 1(4): 6–10. doi:10.1577/1548-8446(1976)001<0006:IFRFFW>2.0.CO;2. https://doi.org/10.1577/1548-8446(1976)001<0006:IFRFFW>2.0.CO;2 (also at https://academic.oup.com/fisheries/article/1/4/6/7884208)
3. Fu, Aihong; Wang, Yi; Ye, Zhaoxia (2020). 'Quantitative Determination of Some Parameters in the Tennant Method and Its Application to Sustainability: A Case Study of the Yarkand River, Xinjiang, China'. Sustainability 12(9): 3699. doi:10.3390/su12093699. https://doi.org/10.3390/su12093699
4. Ma, Lejun; Wang, Huan; Qi, Changjun; Zhang, Xinnan; Zhang, Hanwen (2019). 'Characteristics and Adaptability Assessment of Commonly Used Ecological Flow Methods in Water Storage and Hydropower Projects, the Case of Chinese River Basins'. Water 11(10): 2035. doi:10.3390/w11102035. https://doi.org/10.3390/w11102035
5. Khatar, Behnaz; Shokoohi, Alireza (2025). 'Modification and development of the Tennant method to maintain ecological regime in environmental flow management of rivers'. Sustainable Water Resources Management 11(5): 114. doi:10.1007/s40899-025-01293-6. https://doi.org/10.1007/s40899-025-01293-6
6. Gao, Cheng; Hao, Manqiu; Song, Lianghong; Wang Rong; Shao, Shuaibing; Huang, Ying; Guo, Yufa; Liu, Xueyao (2022). 'Comparative study on the calculation methods of ecological base flow in a mountainous river'. Frontiers in Environmental Science 10: 931844. doi:10.3389/fenvs.2022.931844. https://doi.org/10.3389/fenvs.2022.931844
7. Huang, Yifan; Zhang, Xiang; Xu, Jing; Deng, Liangkun; Li, Yilun (2025). 'Decoding Flow-Ecology Relationships: A Machine learning framework for flow regime Characterization and riparian vegetation prediction'. Ecological Indicators 175: 113517. doi:10.1016/j.ecolind.2025.113517. https://doi.org/10.1016/j.ecolind.2025.113517