Dissolved organic carbon response to hydrological drought characteristics: Based on long-term measurements of headwater streams
Jiefeng Wu,Huaxia Yao,Xing Yuan,Bingqing Lin
DOI: https://doi.org/10.1016/j.watres.2022.118252
IF: 12.8
2022-05-01
Water Research
Abstract:Influence of extreme hydrological events on water quality has been widely concerned. For instance, droughts can inhibit dissolved organic carbon (DOC) exports or imports. However, the response relationship of DOC to hydrological drought characteristics (i.e., duration and severity) requires more in-depth research. We propose an integrated framework for constructing, validating, and applying the response relationship model, and investigate the capability of response model to simulate DOC based on hydrological drought characteristics. Three headwater basins (HP3a, HP4, and HP6), with different drainage areas (9.28-122.80 ha) and long-term (>40 year) observed DOC concentration and hydrometeorological data, in Harp Lake catchment, south-central Ontario, southeastern Canada, are used to demonstrate the proposed framework. Run theory and variable drought thresholds (VDTs) are used to identify hydrological drought characteristics, and DOC during hydrological drought is extracted. Based on the extracted hydrological drought characteristics and DOC for one basin (i.e., HP3a), the response relationship model is constructed and validated, and then applied to other two basins (i.e., HP4 and HP6). Three evaluation indicators: coefficient of determination (R2), root-mean-square-error (RMSE), and mean absolute percentage error (MAPE), are served to test the goodness-of-fit performance of the response relationship model. The results show that (i) annual DOC concentration showed a significant (a=0.01) increasing trend during 1978-2018 in the study basin. (ii) During the hydrological drought, the variation of temperature affected DOC variation indirectly through direct influence on SO4 variation. (iii) The response sensitivity of DOC to hydrologic process with different timescales is varying within a year, namely, there is a larger response sensitivity from March to May than in other months. (iv) DOC during the hydrological drought has a close and regular linear relationship with hydrological drought characteristics, i.e., with the increase of drought duration and severity, DOC concentration also increases. The relationship with drought duration is better than that of severity (R2=0.92 vs 0.35). (v) The response relationship model (autoregressive integrated moving average) can simulate DOC in hydrological drought (R2 ≥ 0.87, RMSE ≤ 0.86, MAPE ≤ 13.69%) at HP3a, and also has good applications at HP4 and HP6 basins. These results provide an improved understanding of DOC-drought relationship, and may support policy makers that look for increased resilience of aquatic ecological security to droughts.
environmental sciences,engineering, environmental,water resources