Climate effects PERSiST

Modelling streamflow impacts on aquatic ecology

Future patterns of streamflow are likely to be a primary driver of aquatic ecology. Using rainflow runoff models such as PERSiST may be one way to evaluate the possible impacts of climate change and land management on the ecology of vulnerable aquatic species. José Ledesma and colleagues published such a study to evaluate future habitat availability the the Montseny brook newt (Calotriton arnoldi). They concluded that future low streamflow conditions will likely pose a severe threat for the survival of the Montseny brook newt but appropriate local management actions including limiting the expansion of holm oak forest may increase the chances for species survival.

The Montseny brook newt is a critically endangered amphibian species which inhabits a small 20 km2 holm oak and beech forest area in northeast Spain. The species can only live in running waters and might be highly vulnerable to hydrological perturbations such as increased drought frequency that could occur under climate and vegetation cover changes. Scenarios describing potential changes in species habitat due to global and local environmental changes can help identify and prioritize the actions needed for its conservation. Based on knowledge of the species and supported by observations, José Ledesma and colleagues proposed daily low and high streamflow event thresholds for the viability of the species. They used PERSiST to simulate changes in the frequency and duration of streamflow events under two climate and four vegetation cover scenarios for near-future (2031–2050) and far-future (2081–2100) periods in a reference catchment. All future scenarios projected a significant decrease in annual streamflow (between 21% and 67%) with respect to the reference period. The frequency and length of low streamflow events was also projected to dramatically increase. In contrast, the risk of catastrophic drift linked to flash floods was projected to decrease. Local hydrologcial effects associated with a potential change in vegetation toward an expansion of holm oak forests will likely be more important than climate changes in determining threshold low flow conditions. This indicates that consideration of both local (potential changes in vegetation) and global (temperature and precipitation) is essential in simulating future aquatic habitats.

Climate effects INCA-C New Papers PERSiST

Modelling DOC in the Canadian Sub-Arctic

Shanta Sharma recently defended her Masters thesis on modelling dissolved organic carbon (DOC) in the Canadian sub-Arctic. This was one of the most northerly applications of the INCA family of models and brought some unique challenges and insights.

The sub-Arctic in Canada and elsewhere is likely to experience hydroclimatic regime change associated with a rapidly changing climate. Simulations of landscape-scale carbon (C) budgets and pollutant transfer are needed by northern managers and stakeholders to understand and mitigate these possible impacts. The project simulated dissolved organic carbon (DOC) fluxes in a hydrologically complex watershed (Baker Creek) in the Northwest Territories. Discharge, DOC concentration, and DOC export were simulated PERSiST and INCA-C. The models were calibrated against available (2012-2016) discharge and DOC concentration data in sub-catchments of Baker Creek. The model successfully reproduced flow (R2: 0.87–0.94; NS: 0.82–0.91) and captured some aspects of DOC concentration dynamics (R2: 0.19–0.31).

Possible future conditions were simulated using two climate scenarios (elevated temperature (T), elevated temperature and precipitation (T+P)), and compared against a baseline scenario. Average discharge is projected to decrease under scenario T (22–27% of baseline) and increase (116–175% of baseline) under the T+P scenario. In this scenario, early winter increases in discharge suggest a change in hydroclimatic regime from nival to combined nival and pluvial. Future DOC fluxes are projected to decrease (24–27% of baseline) under scenario T and increase (64–81% of baseline) in the T+P scenario, with much of the increase in DOC export occurring during early winter. Any future increase in DOC export from Baker Creek may increase the mobility of previously deposited airborne metal contaminants, e.g., arsenic from Giant Mine.

New Papers PERSiST

Hydrological regime shifts in Central European forests?

Petr Kupec and colleagues have published a new paper about possible future hydrological regime shifts in Central European forests. They used long-term data, detailed field measurements from an experimental forest catchent and PERSiST modelling, we show that there is a prolonged and persistent decline in annual runoff:precipitation ratios. This decline is most likely linked to the longer growing seasons associated with global warming. They performed a long term (1950–2018) water balance simulation for a Czech upland forest headwater catchment calibrated against measured streamflow and transpiration from deciduous and coniferous stands. Their simulations were corroborated by long-term (1965–2018) borehole measurements and historical drought reports. A regime shift from positive to negative catchment water balances likely occurred in the early part of this century. Since 2007, annual runoff:precipitation ratios have been below the long-term average. Notably, annual average temperatures have increased, but there have been no notable long term trends in precipitation. Since 1980, there has been a pronounced April warming, likely leading to earlier leaf out and higher annual transpiration, making water unavailable for runoff generation and/or soil moisture recharge. Their results suggest a regime shift due to second order effects of climate change where increased transpiration associated with a longer growing season leads to a shift from light to water limitation in central European forests. If their finding can be generalized, it will require new approaches to managing forests where water limitation has previously not been a problem.


Nitrate leaching under climate change

Climate change may alter the services ecosystems provide by changing ecosystem functioning. As ecosystems can also resist environmental perturbations, it is crucial to consider the different processes that influence resilience.  “Climate proofing” can identify potential climate-related threats to ongoing delivery of ecosystem services. Dr. Katri Rankinen and colleagues have published a new study modelling the potential for increased nitrate(NO3) concentrations in drinking water due to climate change. They analyzed catchment-scale changes in ecosystem services connected to water purification in southern Finland by combining climate change scenarios with process-based forest growth (PREBAS) and eco-hydrological (PERSiST and INCA) models. By using the aforementioned model chain, they improved traditional model calibration by including timing of forest phenology and duration of the snow-covered period from networks of cameras and satellite data. They upscaled the combined modelling results with scenarios of population growth to produce vulnerability maps. Their results show that boreal ecosystems seemed to be strongly buffered against increased NO3 leaching by a combination of increases in evapotranspiration and vegetation NO3 uptake. Societal vulnerability varied greatly between scenarios and municipalities. The most vulnerable areas were agricultural regions on permeable soil types.

New Papers PERSiST

Declining streamflow in central European forests

Dr. Jan Deutscher and colleagues present a new study modelling streamflow decline in the Central European uplands. This study is timely as in recent decades the effects of global climate change have caused a continuous drying out of temperate landscapes. In Czech forests, this drying out has been manifested as a visible decrease in streamflow. Dr. Deutscher and colleagues address questions related to the severity of the streamflow decrease and attempt to identify its main causes. They base their analysis on daily streamflow, temperature, and precipitation data measured during five years (1/11/2014 to 31/10/2019) in a spruce-dominated temperate upland catchment located in the Czech Republic. Streamflow values were modeled in with PERSiST using precipitation and temperature values obtained from the observational E-OBS gridded dataset and calibrated against in situ measured discharge. Their modeling results show a greater than 65% decline in streamflow during the five-year study period at the Křtiny experimental catchment. This trend is most likely caused by increasing temperature. They found a strong relationship between increasing temperature and decreasing discharge during the growing seasons, which can be simplified to an increasing trend in the mean daily temperature of 0.1o C per season, effectively causing a decreasing trend in the discharge of −10% per season regardless of the increasing precipitation during the period.


Spatially explicit, landscape-scale modelling of GHG sources and sinks

Maria Holmberg and colleagues present an approach to collate spatially explicit estimated fluxes of GHGs (carbon dioxide, methane and nitrous oxide) for the main land use sectors in the landscape, and show how these fluxes can be aggregated to calculate net emissions of an entire region. They used INCA-C and PERSiST to estimate the flux of organic carbon from terrestrial ecosystems to lakes and rivers.

They developed and tested the approach in a large river basin in Finland, providing information from intensively studied eLTER research sites. To evaluate the full GHG balance, they included fluxes from natural ecosystems (lakes, rivers, and undrained mires) together with anthropogenic fluxes from agriculture and forestry. They quantified fluxes using an anthropogenic emissions model (FRES), a forest growth and carbon balance model (PREBAS), and literature values for emissions from lakes, rivers, undrained mires, peat extraction sites and cropland. Spatial data sources included CORINE land use data, soil map, lake and river shorelines, national forest inventory data, and statistical data on anthropogenic activities. Emission uncertainties were evaluated with Monte Carlo simulations. They summed the vertical fluxes of spatially explicit net emissions, disregarding the impact of lateral fluxes from terrestrial to aquatic ecosystems on the vertical fluxes.

Their model results showed that artificial surfaces were the most emission intensive land-cover class while lakes and rivers were about as emission intensive as arable land. Forests were the dominant land cover in the region (66%). The forest C sink decreased total emissions for the region by 72%. The region’s net emissions amounted to 4.37 ± 1.43 Tg CO2-eq yr-1, corresponding to a net emission intensity 0.16 Gg CO2-eq km-2 yr-1, and estimated per capita net emissions of 5.6 Mg CO2-eq yr-1. Using INCA-C and PERSiST, the amount of organic C leaching from mires, cropland, and forests to the watercourses was estimated to correspond to about 10% of the CO2 and CH4 emissions from land to air.

Although the landscape approach developed by Dr. Holmberg and colleagues opens opportunities to examine the sensitivities of important GHG fluxes to changes in land use and climate, management actions, and mitigation of anthropogenic emissions, there is still a need to extend the work to a fully integrated regional GHG budget, accounting for all lateral fluxes of C- and N-containing compounds.