My name’s John Ossanna. I’d like to welcome you to our webinar series, held in partnership with the US Geological Survey’s National Climate Adaptation Science Center. Today’s webinar is titled “Assessing Impact of Future Climate and Introduced Species on Hawaii’s Aquatic Ecosystems.” We’re excited to have Yin-Phan Tsang and Hannah Clilverd, from the University of Hawaii at Mãnoa with us today. To introduce our presenters, we have Janet Cushing, who is the acting director for the Pacific Islands Climate Adaptation Science Center. Janet, you have the floor. Thank you very much. On behalf of the National Climate Adaptation Science Center and the Pacific Islands Climate Adaptation Science Center, I am honored to introduce our two speakers for today. Yin-Phan Tsang is an assistant professor in natural resources and environmental management at the University of Hawaii at Mãnoa. She’s particularly interested in hydrological processes, hydrological modeling, and ecohydrological implications and applications to aquatic ecosystems across multiple spatial scales. Our second speaker, Hannah Clilverd, received her PhD in geography from the University College London in the United Kingdom, based on ecohydrological monitoring and modeling of river floodplain restoration. She’s currently a post-doctoral researcher in the Department of Natural Resources and Environmental Management at the University of Hawaii at Mãnoa. With that, I turn it over to our speakers. Thank you so much. I’d just want to say thank you very much for that introduction and also for this opportunity to present. Before I get into the talk, I would also like to thank our funder, the USGS National Climate Adaption Science Center. I’d also like to acknowledge my co-authors, Dr. Yin-Phan Tsang, Dana Infante, Abigail Lynch, and Strauch. Today, I’m going to cover a bit of background on streams in Hawaii, and in particular, the importance of upstream-downstream connectivity to fish. The talk will focus on two main topics, how streams are changing in Hawaii and then the potential impact on native fish. The climate change is expected to directly impact the structure and functioning of tropical streams by altering the hydrological, thermal, and chemical conditions. Stream systems in Hawaii are particularly important because they provide more than 50 percent of irrigation water to the island. They’re very important for agriculture. They host endemic stream fauna, and they also, amongst other things, influence the condition of natural and coastal habitats in terms of sediment and nutrient lakes. It’s important to understand how climate change affects these resources. In recent decades, mean surface temperature in Hawaii has increased rapidly, consistent with the global trend. We know that this warming is already having documented impacts on the hydrological cycle. Recent studies have shown a decrease in the total amount of rainfall as well as the intensity of rainfall, and an increase in droughts occurring. This is likely to impact groundwater and recharge as well as thin streams. Indeed, we have been in a downward trend in base flow from 1930 to 2008, which has been what’s in literature. If this continues into the future, this has potentially serious consequences for streamflows and the distribution and abundance of native organisms. Just to give you an idea of what we can expect in the future, statistical downscaling and dynamical downscaling of global climate model output indicates that we’re likely to see strong dipolar rainfall patterns, with the wetter windward side of the islands becoming more wet and the leeward drier areas of the islands becoming drier in the future. Streams in Hawaii are very splashy as shown here in this hydrograph in the top right of the stream in Maui. This is driven by the steep topography, the short peak to ocean distances, the low water stream systems, and then intense tropical rainfall on top of that. Rainfall can exceed more than 10, 000 millimeters per year in some windward areas, which is just remarkable, It makes Hawaii one of the wettest places on Earth. This is a map showing the spatial distribution of perennial and intermittent flow across the main Hawaiian Islands. The perennial flow is shown in dark blue, and the intermittent flow is shown in light blue. You can see the perennial streams typically occurs on the windward sides of the higher elevation, geologically younger islands, and intermittent flow and low rainfall occurs on the leeward areas. Freshwater resources in leeward regions can be quite limited, and such. There’s quite an extensive ditch network across Hawaii, that’s used to divert surface water from the wetter areas to the drier areas to meet agricultural water demand. This has, as you can imagine, resulted in water disputes where in-stream flow standards have been impacted. The rather small number of taxa that comprise the native fauna of Hawaiian streams, and these are endemic to the islands. These include five species of amphidromous fishes, two species of amphidromous shrimp, and two species of amphidromous snails. These are distributed longitudinally, based on their waterfall-climbing ability. You can see one of the better climbers here, shown in this video that’s and scaling a waterfall there. These migratory species, they lay their eggs in the streams, which then hatch and are washed out to the sea. Then they disperse and spend a marine larval phase feeding, before returning back to freshwater as juveniles. That’s where they spend the majority of their life, in freshwater. Unobstructed passage is very important for these species. One, for the initial dispersal of the larvae, but then also for the recruitment of the juveniles later on. I’m just going to go over quickly some of the objectives of this study. Firstly, to examine the hydrographic record for evidence of changes in flow regimes. This assessment has been on the observed trends in stream flow over the last 50 years. Secondly, to determine regions where streams that provide effusia for native species may be more susceptible or resilient to the impacts of climate change. This analysis has used species data for the state. The third objective is to determine future streamflow dynamic patterns across the Hawaiian Islands through modeling simulation. This is something that we’re working on now. I’m going to get into some of the methods here, so in order to climate-driven changes in streamflow, we needed to identify streams that were unimpacted by anthropogenic incidences, so namely, surface water diversion. You can see a picture here of the surface water diversion taking all of the base flow from a stream. We examined 390 USGS gauged reaches. We also looked for streams that have a continuous data set and also, a long data set, so we needed 50 years of data. Based on all these criteria, we were able to identify 23 unregulated streams. These are, quite luckily, distributed quite evenly across the Hawaiian Islands. They encompassed a range of hydrological and climatological conditions, although we were hoping that we would’ve gotten more stream gages in leeward areas, but just wasn’t available. To look at the trends, we used a non- parametric sen slope estimator and a Mann-Kendall test for significance. These tests are recommended for analyzing environmental time series data. They’ve been widely used for hydrological data. For looking at the distribution of native species across Hawaii, we have presence-absence data from nine taxa so approximately 10 percent of perennial stream reaches. This included just 12 of our study sites, 2 of which actually recorded no species. This is a wonderful data set, and we’re very grateful to the Hawaii Department of Aquatic Resources and the Hawaii Fish Habitat Partnership to use this data. This was used in an elimination analysis, some of which I’ll show you today just the species associations using the Kenoco Program. On to the hydrological modeling. We’ve identified streams across the state based on leeward and windward areas, again, using more natural streams but, at least, not impacted by surface water diversion or ground water pumping upstream of the gauging station that we’re using for the model. At the moment, we’re focusing on Oahu. We’ve identified three watersheds, that have also importance for native fish. We have data available for them. I just highlighted Punaluu watershed and the SWAT model that we’ve put together. SWAT stands for Soil Water Assessment Tool, and this is a watershed scale model. It’s specifically based on three distributors. It requires inputs of climate data, land use, soils and floods. We’re actually very lucky that Hawaii has this really wonderful data set of daily rainfall grids for the state, a really fine revolution which has helped us with the model development in capturing those high spatial variability in rainfall inputs with these very small watersheds. I’m going to go into the results now. This is a look at the spatial patterns for annual streamflow trends from 1987 to 2016. I’ve got baseflow which is your low flow conditions in the top panel and runoff which is your strong flow conditions in the bottom panel. This is superimposed on mean annual rainfall from the rainfall outlets. The red triangles indicate declining streamflow and blue triangles indicate increasing streamflow. Anywhere you see a black dot, that’s a statistically significant trend. What I want you to take away from this is that there is variability in the trends across the islands. Overall, we are seeing the annual baseflows and runoffs have declined across the Hawaiian Islands, on average 11 and 8 percent respectively, which is indicating a reduction in water availability in most of the study streams. We actually see stronger trends in baseflow. We could see a significant downward trend in 60 percent streams for baseflow and 22 percent for runoff. We found that Hawaii Island, in particular, is exhibiting strong, significant, negative trends in baseflow and in runoff. Significant declines in baseflow also occurred in northeast Maui, and in all but one study stream on Kauai. As I just said, these patterns are not homogeneous, so we see that declines in baseflow were not significant in the majority of streams on Oahu, which had some of the lowest detectable baseflow and runoff trends of all the islands. We’re also interested in how the trends changed through time, with the length and starting date of the analysis. Here’s the running trend analysis results. There’s a lot of data in this figure, so I’m going to walk you through it, and I’m going to show two slides like this. Along the x-axis, we have the start year of the analysis. On the y-axis, we have the duration of the analysis. As you move along the x-axis, your duration of the analysis decreases. We have baseflow on the left and runoff on the right. Each square represents one trend analysis. Anything in red indicates declining streamflow, and anything in blue indicates increasing streamflow. Then, I’ve outlined boxes that show a statistically significant trend. What this analysis tells us is that it’s not just a simple story of decreasing trend. Yes, we do see decreases in baseflows, and they do dominate the trends and strengthen with time. We can see these 20 to 30-year patterns between increasing flow and decreasing flow. These broadly line up with the negative and positive Pacific Decadal Oscillation phases. However, the recent decline is not constrained to the positive PDOs, so it does actually continue downward. That’s something to keep in mind as you look at the next figure. There does seem to be some inter-decadal variability that is driving these trends. Downward trends were most pronounced on the Big Island, particularly in dry areas. You can see in this figure that it’s very red. We found that streams on the younger islands, so Hawaii Island and Maui, which have less developed soils and young, permeable volcanic substrates, these islands are responding rapidly to changing rainfall. Consequently, they’re exhibiting this particular vulnerability to drying conditions. What we’re concerned about, if these strong declines continue into the future, they may end up reducing water availability during lower rainfall. Another metric that we looked at was the number of no-flow days, which is important to understand for upstream- downstream connectivity and the maintenance of fish habitats. If I draw your attention to the top right-hand figure, three of our study streams showed intermittent flow. Of particular interest is Kaluanui Stream, which is shown in blue. This is a pristine watershed that also has a diversity of native species and shows a degree of flow intermittence so it’s really interesting. That tells that these native species can or are tolerable or are not harmed by the small amount of flow intermittence. Another stream to point out is Hakalau Stream which is on windward Hawaii. This stream, shown in red, has begun to exhibit no-flow days since 2014. This is unprecedented in the 50-year record. It’s obviously something of concern and requires further investigation. Just as a point of comparison, I’ve included a stream in the bottom panel that has been impacted by groundwater pumping. This is a dry area of Oahu. You can note the difference in scale on the y- axis and the additional groundwater well was bottomed or mined in 1990. That had a substantial impact on the number of no-flow days. This stream actually used to flow perennially to the ocean but no longer or, at least, during this period shown here, it was no longer flowing to the ocean. Incidentally, this stream has no native species present. I just thought I’d give you an indication of the impact the groundwater pumping can have on baseflow, given that I’ve already showed you the impact of some of the surface water diversion. On to the biological data. I’m just going to show one of results which is from a detrended correspondence analysis. The open circles in the figure represent the stream reaches which we have data for. Then the gray circles highlight the streams from the climate analysis for which we have hydrological data. Then the triangles represent the stream taxa. You can see here that overall we found two distinct groups of species associations among the sample reaches which as you would expect indicates a degree of habitat separation between species that can climb the waterfalls, for example, Lentipes and , and the species that cannot climb on the left, Stenogobius . This south access explains about 30 percent of the total species’ variability and does appear to follow a gradient in climbing ability. The sites in the middle have higher presence of different-natured species. Our trend analysis sites, in particular, carry new streams on windward Oahu stands out. The sites that fall out in the middle are where niche overlaps occur. Interestingly, these sites vary considerably in terms of their hydrology. Some then have very high baseflow, whereas others are more flashy. They indicate a similarity in community composition which reflects the adaptability of the native Hawaiian fish to variable hydrological conditions in Hawaii. Our goal now is to further constrain this analysis with flow data which, unfortunately, we don’t have for most of the study sites because they’re engaged. What we’re working on is a regression relationship between rainfall and flow. It can use rainfall as a proxy for flow metrics. I’m going to go on to the last bit here which is the modeling part of the study. I’m just going to show you Punaluu Watershed. This is a really interesting watershed. You can see that in this figure here the time series of the top panel that this stream has quite a high baseflow which is caused by the presence of dike complexes upstream which are basically low permeability, volcanic complexes that parch the water table. Because of this good water source, the low lying areas have actually been historically used for farming. Today, the area is mostly used for small-scale, diversified agriculture. In 2008, which you can see down here, they installed a new ditch which is, and I don’t have a picture of it, but it is a pipe intake. It has substantially reduced the amount of water that is taken out of the stream. You can see here for 2014, for which we have data, the percentage baseflow diverted has been substantially reduced. This is one of the best examples of an improved fish-friendly ditch intake design and management that we have in Hawaii. I think that there’s some capability of the intake pipe to be closed at night to aid the downstream transfer of native larvae so that’s quite interesting. For the model, we’ve used 10 years of stream discharge data to calibrate and validate the model. These years were chosen to capture range of hydrological conditions that incorporated El Nino and La Nina events. This is a quick look at the model set. You can see that we’ve captured the timing of flow events very well as well as the magnitude of baseflow, in particular, which was our focus for modeling fish habitat. We have a match, cyclic statistic in the region of about .8 which indicates very good agreement. Now that we’re happy with our model, we start to perturb it with climate projections from 2015. We’re using seasonal rainfall and temperature anomalies for achievable, minimum and business- as-usual which are the RCP 4.5 and 8.5. Our current analysis suggests that this watershed might be fairly resilient of the projected changes in climate. You’re looking at baseflow and changes in flow intermittence. This is essentially very important for providing refugia for native species having high conservation importance on Oahu, given that so much of the island is highly urbanized and all of the problems are associated with that. It’d be really interesting actually to look at the streams that we’re currently simulating streamflow for which is next door to Punaluu. This stream, which I showed you, the flow in an intermittent state for earlier already exhibits the degree of flow intermittence so we’ll see how this stream responds to the climate scenario. With that, I’m just going to wrap things up with a couple of take-home messages. The streamflow has decreased in many streams across the islands. Of most concern throughout streams in drier areas that it may become more intermittent. While we can’t really be certain that the patterns in this study are not attributed to a degree of natural variability associated with the PGO, what we can be certain of is that the declines we see in baseflow from water diversions far exceed the percent, the kinds that were observed in the study. If we do want to buffer these systems against climate change, then we do need streamflow protection and restoration that’s going to return flows and improve habitat quality for key native species. With that, I’ll finish up, and I think I’m going to pass on to Yin-Phan. Thank you everybody for giving us this opportunity to speak. Hannah gave us a really good segue to a broad perspective of the streamflow for Hawaii. We see the decreasing trends throughout most of the area, most of the islands, and try to understand what the implication to our native species. Following that, we are going to talk about something a little bit different. We’re going to talk about our introduced species, in our Hawaii’s stream ecosystem. This work is being done by my student, Brendan Martin, and with a lot of collaborator, including Hannah, and Ralph Tingley, Dana Infante, Kyle Herreman, and Glenn Higashi. What is introduced species? Introduced species are living… What that means is that they live outside of their native range. They have been called introduced species, aliens, exotic, non-native, and it has been a major threat to our freshwater ecosystem. Excuse me. It’s advancing on its own. I’m going to try it again. Specifically, the establishment of the introduced species is a lot of times caused a loss of our native species, because most of the time, these introduced species have higher tolerance to harsh environment. Is mostly competitive and have a wider food preference. Sometimes fast- growing, and they have low age of the maturity. The introduction of these species to Hawaii is due to the need of aquaculture or mosquito control, and sometimes sport fishing, or is ornamental. The result of it often becomes like the changing of the stream habitat, physically and chemically. Given that we know that Hawaii has a really unique stream ecosystem, we also know that the natural environment in Hawaii has a really large gradient in terms of their natural factors, as well as their human disturbance factors. At the same time, we have really relatively fewer native species. We like to understand that how does introduced species, what’s their habitat usage, and what’s their distribution, for our Hawaii’s streams and its ecosystem. Before we go into that, we know that stream is really a complex ecosystem, especially it’s in hierarchical structure. For example, the basin, the reach, the channel unit, they are in very different spatial units, and at the same time, they affect each other. We know the habitat, the in-stream channel units, and micro-habitat, is very important. It’s complex and strongly influence our local species. At the same time, those landscape-scale basin and reaches is also affecting our habitat, and in turn affecting our species. At the same time, the human disturbance is affecting across different scale for our stream ecosystem. For this study, our objective has three parts. The first one, we want to try to characterize the association between the species and the in-stream habitat attribute. We call it habitat assessment. For the second part, we want to characterize the association between the species and the natural and anthropogenic landscape factors. We call it landscape assessment. For the last part, we’re going to model the suitable habitat for the introduced species using the landscape scale environmental factor, and we call it the modeling assessment. Here is the study area we have, the five main Hawaiian Islands. As you can see, those blue lines are indicating the perennial streams. The orange lines are those intermittent streams. My apology for the advancing pictures of my slides. I’m not sure what’s going on. For the biological data though, we do have two different kind of biological data. One is abundance, one is the presence- absence. Both of them are conducted using snorkeling visual survey in the stream reach. For the abundance data set, it is a stationery survey at the one times one meter point quadrant survey. We have almost 8,000 number of the survey. The time was surveyed between 1989 to 2010. We have another type of survey, which is the presence- absence survey. It’s a moving survey across the 50 to 100 meter of stream channel. We have 466 of the survey data. It’s spanning from 1970 to 2014. Both of them have, at least, 41 introduced species in those survey data sets. Here is the spatial distribution of our survey location. The green ones are those abundance survey location, and the purple ones are those presence-absence survey location. You can see that they are distributed throughout most of the perennial stream and some are in the intermittent stream. For the first habitat assessment, again, our objective is trying to characterize the association between the species and the in- stream habitat. We use the abundance survey in trying to associate them with those in- stream survey attributes. We conduct the correspondent analysis, the CCA. Then we use the zero inflect model to fit our data as well for those taxa or species have, at least, 10 survey occurrence. We use the distance to grow an elevation to predict the zero, and we use reduction using the Akaike Information Criterion to reduce some of the parameters. The reason that we need to use the zero inflect model specifically because this is using for some of the data that they have is not fitting to the traditional normal distribution of data, especially have a lot of zero data. This zero, some of them they are true zero. What that means is that they are not suitable habitat with the species that we’re looking at. But they are some false zeros. For our cases, those are the location that the taxa haven’t been introduced to those locations. By using this method, we can model the true zero and the false zero. Then along with those abundant data, the taxa count, we can do a better prediction of the habitat simulation. Here is our first part of the result. This is our CCA result. The arrow showing the numerical variables and the direction is point to the increasing direction. The red triangles are those of categorical variable. In our case, it’s mostly the habitat types. The blue triangles are the species distribution. As you can see, the dominant factor in all our data set that describing our introduced species distribution are temperature and the depth. Specifically, you can see that the American bullfrog, aquaphilia, and tilapia that is in a higher temperature habitat while smallmouth bass and the red swamp crayfish is in the deeper upstream habitat. At the same time, you can see that these are the group that associate with the habitat that have the gravel or the sand habitat substrate. At the same time, we find that there are a group of species that actually at the center which are signaling they might be both general habitat user. Here is our zero inflect model result, and the that is indicating the significant level. We first look at the zero modeling using the row and the elevation. We found that the distance to the row is definitely signaling how much the zero is happening. The further from the row, we have more zero of the species present. The association between the zero and the elevation is mixed. For the habitat model part of the pond and the red shrimp crayfish and are more associated with the mid or large substrate. For the American bullfrog, we see that they are more associated with the , which is the decayed organic matters. This is our habitat assessment result. We move on to the landscape assessment. Again, our objective is trying to characterize the association between the species and the natural and anthropogenic landscape factor. To do that, we first need to link up with our spatial framework. Our spatial framework is based on this framework that has been established previously for Hawaii species habitat assessment. Our basic spatial unit is here, the pink reach. There are three spatial scales that are associated with this reach. The first one is local catchment, which is that yellow area. The area drains directly to this pink reach. There’s the upper catchment, which is all the area from the stream network that drains into this specific reach. It’s including the green and the yellow area. Finally, we have this pink or reddish area. Those are indicating the downstream main channel catchment. All the natural and anthropogenic factor is aggregate into the three spatial scales. For our biological data, we link all the seven locations to individual reach. When a species is shown in the survey which is called a present. If the species is not occurrence within that survey at that given reach, we call it absent. We conduct the forward selection CCA to select the landscape factors. Finally, we do indicate a species analysis trying to associate the landscape factor with our species. Specifically, here’s an illustration. If we have environmental gradient along this axis, the species will show from the absent and become present. We call this is a positive increasing or positive changing point at that given environmental value. On the other hand, if the species along this environmental gradient is from present to absent, we call this as a negative changing point at the environmental value. Here comes our result of, first of all, our forward selection CCA with the p- value locked in .01. You can see that the natural factor and the anthropogenic factor are list here. The natural factor, including the downstream channel slope, the upstream mean annual rainfall, etc. Then we have anthropogenic factor that have downstream row length and upstream pipeline length. This is our forward selection CCA. For our indicator species analysis result, here are the result of five natural factor. Let me walk you through how we read this. On the bottom of the x-axis is showing an example of natural gradient, which is the upstream mean annual rainfall. You can see different color of dots on the top indicating the species that is increasing occurrence or decreasing occurrence along the gradient. The red dot indicating increasing, the black dot indicating decreasing. Then the size of the dot indicating the significance of the changing point. The line is indicating the 95 percent of competent level of bootstrap analysis. You can see for this one, most of the species is showing a decreasing occurrence along with the upstream mean annual rainfall. The exception is the Tahitian prawn. It’s actually showing an increasing occurrence at about 2,400 mean annual rainfall. In general, if you look at the natural factor, most species are showing decreasing trend along the local elevation, the upstream mean annual rainfall and downstream flow that is increasing along the gradient of the local mean annual air temperature and the downstream level. This is the natural factor. The following are the three anthropogenic factors that we want to highlight. As you can see, most of the species are showing increasing along with the anthropogenic gradient, including the downstream agriculture, the upstream population density, and the downstream row density. Again, the exception is the Tahitian prawn. The Tahitian prawn has behaved really differently from other introduced species. I’ve got to move on to the third part of our analysis. Again, this objective is trying to model the suitable habitat for our introduced species. The method that we like to use is trying to model their suitable distribution for these non- native species. Again, we are modeling their non-native range. The advantage of it is that we are incorporating some of the novel environment, especially where in Hawaii there are other factors that is different from its native range. The con of this modeling approach is that the species might not have been introduced or reached to all the potential suitable habitat. I want you to keep in mind that there is a caveat here that our predicted suitability of the habitat may not be adequate, or it might be narrower than the actual suitability. The modeling approach we use is the Booth regression. Then we use the random template cross- validation to ensure the modeling quality. We select the text with, at least, channel more occurrence throughout our data set so there is a subset of the species that we are modeling. We use a selection of the predictor that’s from literature, three of them — the reach elevation, the channel flow, the catchment area — and we select from the four selection CCA. We select nine predictor of the landscape factors that are listed here. Eventually, we use this cross-validation area under the curve AUC method to evaluate our model quality. Here is the results. Based on our model results, we found out there are some species model that is showing good model, which have, at least, AUC larger than .9. That’s the green turtle and the tilapia. We have satisfactory model that have AUC larger than . 8, including Tahitian prawn, the western mosquitofish, and cilius. We also have some poor model which we won’t be showing here, but there are some inadequate data, that we couldn’t do a good adjustment of the model. We found that most of the important predictor, including the downstream slope, the local impervious surfaces, upstream mean annual rainfall, and reach elevation, and specifically that we compared the predict reach through the , along with our observed occurrence of the species. As you can see it here, the Tahitian prawn that is showing way more suitable habitat than what we have observed. In the second place, is the western mosquitofish. Now, here is the spatial distribution of our model’s result of tilapia, and I’m going to walk you through this. From the red, orange, green, to blue, it’s indicating that for high to low stability of the species, and specific for tilapia, the important factors are downstream slope and the local impervious surfaces. As you can see, most of the red area is near the coast, along the shoreline, especially in Oahu and Kauai. Here is another map for Tahitian prawns, and for the same legend, then the important factor for Tahitian prawn is elevation and the upstream mean annual rainfall. You can see that a lot of the red or orange area is showing at the higher elevation in some of the Oahu, Kauai, and along the shoreline of Hawaii and Maui. Just to sum it up our study, first of all, we found that the water stats and the temperature are explaining most of the variation in our taxa distribution for our introduced species. The species is generally associated with a channel unit or the habitat types. American bullfrog, tilapia, and crustacean are showing very distinct habitat usage. For our last , the rainfall and slope is limiting our introduced species distribution. For anthropogenic factor, population and row density is highly associated with our introduced species. Elevation might have to represent a mixed signal between the natural factor and anthropogenic disturbance. For the modelling assessment, we found that the Tahitian prawn exerts the greatest amount of super reach for the entire state of Hawaii. We found that the tilapia stability were more limited to the lower elevation near the coast area, especially for Kauai and Oahu. Just to give you a preview about what we just recently accomplished, we was trying to combat the waterfall assessment to evaluate if waterfall is acting as a natural barrier for the distribution of our introduced species. We used the Google Earth to identify the waterfall location. We measured the waterfall height. Then we tried to use a landscape approach to assess the…I don’t know why it keeps working this way. We trying to assess the process of introduced species was showing because of the human introduction or it was barrier. The waterfall is become a barrier for them to climb up. This has been a write-up and being submitted to a landscape approach book chapter to American Fisheries Society. Stay tuned and we will have more to report. Sorry about my crazy slide advancing, but I really like to thank our funders and our collaborators. This is a lot of work but we got a lot of input and help to accomplish this work. Thank you. I will take any question if you have any. Thank you for both of your presentations. We do have a couple of questions already. Our first question comes from Ethan. How were the streamline data generated, field map or flow model? The streamline was actually starting with the NHD streamline that is a national product, but then it has to be modified based on local input. Especially, this was previous work that had been done to assess the stream habitat for Hawaii. I would say that yes, it’s a combination of both. It’s not just a model GIS work, but also with local input. Our next question, this one is for Hannah. Is there a good understanding of the minimum flow requirements to support the various fish species? Even with the current declining trends in average annual flow given the flashness of the streams, what is the likelihood that we will cross that threshold of going below minimum flows at the time for the fish that are migrating inland? I’m sorry about that, Hannah. Can you hear me? Yes. That’s a really good question. Folks are looking into in-stream flow standards. I don’t think we have a particular threshold that we can say that these fish are excluded. We know, as I showed, that some degree of flow intermittence is tolerable by these species. We think it could also be potentially beneficial and to a certain extent perhaps following a flash flood event if that doesn’t move out invasive species from a stream. Major species are able to re-inhabit the area and a small amount of flow intermittence might be beneficial to them. This is something that probably needs further investigation to see at what point does intermittent flow severely interrupt the passage of these species. That’s something I’d definitely like to look into more. There’s another question we have from Springer K. There has been some interest in removing invasive Albizia trees. I hope I said that right. Where they overtop streams in the Eastern Hawaiian Islands. This could reduce inputs and reduce spread of the invasive tree, but would expose streams to bright sunlight. Good idea, bad idea, thoughts? I have to say I’ve never really thought about that in terms of temperature. I guess I would assume if they’re an invasive species, then you would prefer a native assemblage in the upper watershed. Most of the studies that I’ve looked at to do with Albizia have been associated with the impact from flooding. Yin-Phan, do you know anything about whether they would substantially impact stream temperatures? I’m not recalling any studies that have been talking about the temperature, which is a really interesting factor, especially when I do see that increasing temperature in Hawaii. How much that is really affecting our stream, that’s the question. I don’t recall any temperature study along with our native species yet, but that’s a good question. We should take a look. There are concerns with non- native species impacting the hydrology in terms of the impact on evapotranspiration and things like that. Maybe, that could be something that is important. Certainly, we are aiming to restore our native species to the whole upper watershed but to the upper areas unless impacted by people. I’m impressed that the is going so well at the East High Mountain Hawaii. At this time, I would like to thank both Hannah and Yin-Phan for their presentation. I’d like to thank USGS for their continued support of this webinar series. Nothing left to say. I would like to say once again thank you all and have a nice day. Thank you very much. Thank you.