How to estimate sediment from Total Suspended Solids?

[Liem, Nguyen Duy]

Dear all,

I am doing a research related to modeling sediment using MUSLE in SWAT model. After finishing run model, I would like to calibrate simulated sediment using observed data. Unfortunately, I just have observed data of Total Suspended Solids. 

I am wonder how to estimate sediment from Total Suspended Solids? What should i do to overcome the problem?

Thanks for your reading and looking forward to hearing from you!

Liem

--

Nguyễn Duy Liêm

Bộ môn Tài nguyên & GIS,

Khoa Môi trường & Tài nguyên,

Trường Đại học Nông Lâm TP. HCM.
Địa chỉ: P. Linh Trung, Q. Thủ Đức, Tp. Hồ Chí Minh, Việt Nam

Nguyen Duy Liem

Department of Natural Resources and GIS, 

Faculty of Environment and Natural Resources

Nong Lam University- Ho Chi Minh City.
Address: Linh Trung Ward, Thu Duc District, Ho Chi Minh City, Vietnam.

Email: nguyen...@gmail.com

          nguyen...@yahoo.com

Mobile: (+84) 0983-613-551

Website: http://nguyenduyliemgis.wordpress.com/

Linkedin: https://www.linkedin.com/pub/liem-nguyen-duy/29/9bb/7aa

ResearchGate: https://www.researchgate.net/profile/Liem_Nguyen_Duy

Academia.edu: https://hcmuaf.academia.edu/LiemNguyenDuy

Facebook: https://www.facebook.com/nguyenduyliem.gis

Youtube: https://www.youtube.com/user/nguyenduy11331

[Liem, Nguyen Duy]

Dear Willem,

Thank you so much.

In my study basin, there is no sediment data except TSS.

When I compare TSS data with simulated sediment data from SWAT, it showed that they are the same trend but sediment data always higher than the other (about 100 times). 

As I know, TSS is a part of sediment, therefore I would like to know there is any way to estimate sediment from TSS or extract TSS from sediment? 

I am looking forward to hearing from you!

Liem

On 6 September 2015 at 06:47, Willem Vervoort <willemv...@gmail.com> wrote:

Liem,

I don't think there is a straight forward relationship. TDS also includes all salts. However, if in your case sediment dominates (say 90%) TDS, you could probably use TDS to estimate sediment.

In my case, sediment would only be a minor component.

Do you have other data, such as EC, or other estimates of salinity, so you can subtract the soluble salts?

Willem

--
You received this message because you are subscribed to the Google Groups "ArcSWAT" group.
To unsubscribe from this group and stop receiving emails from it, send an email to arcswat+u...@googlegroups.com.
To post to this group, send email to arc...@googlegroups.com.
Visit this group at http://groups.google.com/group/arcswat.
For more options, visit https://groups.google.com/d/optout.

--

http://blogs.usyd.edu.au/waterhydrosu/

[Iman Fazeli]

Dear Willem and Liem,

The sediment output in SWAT is "tons" in output.rch, it means SWAT show you "tons" in each subbasin.

About Liem's question, As far as I know there is not a logical way to use TSS as sediment in SWAT-CUP unless you have EC and also salinity parameters separately. 

I would like know other opinions about this good question. Maybe someone suggest a good idea.

Best,

Iman

On Mon, Sep 7, 2015 at 3:18 AM, Willem Vervoort <willemv...@gmail.com> wrote:

My suggestion would be that you overestimate sediment, unless there is a unit conversion. I would think TSS is in mg/L, but I am not sure what the sediment output from SWAT would be (Something per m3?). Because TSS also includes salts, it should be higher than suspended sediment. However, is the SWAT estimate including something like bedload? I am not very familiar with the sediment routine.

Regards

Willem

--

http://blogs.usyd.edu.au/waterhydrosu/

[Liem, Nguyen Duy]

Dear Iman,

Thanks for your comment.

For more clearly, I converted the unit of TSS from g/m3 to metric ton before comparing with sediment output of SWAT.

Liem

[Balaji Naraimhan, PhD , Asst. Prof., IIT Madras]

Liem,

Are you interested in comparing TSS or sediment load?  

Could you indicated what sediment routing equation you are using?

If you use the default sediment routing equation, then the TSS is simply (sediment load/volume of water)

If you use other methods of routing that partitions sediment into their size categories, then TSS would be (silt + clay)/volume of water.

Balaji

[Jim Almendinger]

I agree with Balaji -- I think this is the most straightforward way to interpret the SWAT output for sediment. 

-- I typically try to calibrate on monthly loads rather than concentrations, if only because grab-sample concentrations can be so variable.  On the other hand, it's good to remember that "observed loads" themselves are rather uncertain numbers based on observed flows (fairly reliable) and observed concentrations (potentially highly variable). 

-- There is some question in my mind whether the MUSLE results include all grain sizes or just the suspended fraction.  The USLE results should include all fractions because this equation was based on plot-scale studies where all grain sizes were captured.  But MUSLE was based on (if I recall) small watershed-scale loads for 18 watersheds from the southern USA, and I don't know if bedload was included or not -- and if so, whether it could be partitioned between field and bank sources.  MUSLE loads are almost always less than USLE loads, and I've presumed part of this was because of missing the bedload component, along with other sediment traps along the way between fields and the gauging station.  

-- I presume if you use the default sediment equation in SWAT, which assumes silt-sized particles, and you calibrate to the suspended load, then you simply can't say much (if anything) about bedload.  If you use one of the grain-sized specific methods, then make sure you calibrate only the clay+silt fractions to the observed TSS, as Balaji points out. 

-- Our measurements of TSS involve filtration of a known volume of water through pre-weighed glass-fiber filters, which are then dried at 105 C and re-weighed.  The dissolved mineral component is thus not included (except for just the bit of water that wets the grains on the filter, after filtration).  TSS includes both inorganic particles (usually dominant in agricultural rivers) and organic particles (volatile suspended solids (VSS), which are often algal bodies that can dominate in the outflow from reservoirs and lakes).  If you're most interested in soil erosion of largely inorganic soil (>95%), you probably should subtract out (most of) the VSS if you know it, although I don't know of many who do this. 

-- I will add that in recent calibrations I've been working on in systems with many lakes, the model results are very sensitive to the grain size (D50) I set for each lake, along with NSED values.  The lake D50s had to be set in the fine-silt/clay range for the lake to pass much sediment downstream.  And, NSEDs had to be set low, in the 1 mg/L or less range -- if they're too high, then lakes become sources rather than sinks of sediment, which is not likely to be realistic in most cases. 

-- If your modeled loads are too large, then SWAT is either generating too much sediment in the HRUs (reduce erosivity, slopes, CNs, or increase vegetation cover somehow), or there are sediment traps along the way (Ponds, Wetlands, Reservoirs) that are not being accounted for. 

-- And, be on guard that the channel erosion equations are not unrealistically eroding sediment from the reaches.  You should check to see where all your excess sediment is coming from.  I think SWAT-CHECK will tell if if it's field or bank erosion.  Once you know the source, you can then target the proper parameters in SWAT to tone down the modeled loads. 

Best of luck,

-- Jim

---

From: "Balaji Naraimhan, PhD , Asst. Prof., IIT Madras" <nbalaj...@gmail.com>
To: "SWAT-user" <swat...@googlegroups.com>
Cc: "SWAT-CUP" <swat...@googlegroups.com>, "ArcSWAT" <arc...@googlegroups.com>
Sent: Monday, September 7, 2015 1:42:40 AM
Subject: [ArcSWAT:7552] Re: How to estimate sediment from Total Suspended Solids?

--
You received this message because you are subscribed to the Google Groups "ArcSWAT" group.
To unsubscribe from this group and stop receiving emails from it, send an email to arcswat+u...@googlegroups.com.
To post to this group, send email to arc...@googlegroups.com.
Visit this group at http://groups.google.com/group/arcswat.
For more options, visit https://groups.google.com/d/optout.

---

Spam
Not spam
Forget previous vote

--

Dr. James E. Almendinger
St. Croix Watershed Research Station
Science Museum of Minnesota
16910 152nd St N
Marine on St. Croix, MN  55047
tel: 651-433-5953 ext 19

[fatemeh babakhani]

Hello Jim

The 5th paragraph of your email, helped me a lot. But something still is not clear to me, I really appreciate if you give me your thoughts about that.

When I set D50=10 (Micro-meter), and NSED=1 (ppm), the sediment accumulation rate will be negative, on the other hand, the reservoir or lake will be the source of sediment, in 2 years (1988, and 1990). I tried smaller NSED, but it didn't work. I should mention that in these 2 years, the flow in to the reservoir is the highest between the simulation period (from 1980 to 1999).  

I expected in the flood events sediment accumulation rate increases, but my SWAT results doesn't show that! I think in the flood events, sediment load has been washed out from the reservoir!

Have you ever seen any reservoir or lake that have the same issue like my reservoir. 

Thanks so much!

Fatemeh

[Jim Almendinger]

Fatemeh --

No, I agree that your reservoir should be trapping sediment, especially during high-flow years.  I'm guessing that the only way to solve the problem is to increase the D50 of the particles entering the reservoir, causing them to settle out more rapidly.  The problem with that is that it may mess up your sedimentation in other years.  Perhaps the bigger problem is that the actual D50 of incoming particles probably changes from year to year with the incoming flow -- the greater the flow, the greater the particle size, I would think.  Ideally, of course, SWAT would track particles sizes from field to stream to reservoir (and perhaps some of the newer algorithms do, but I haven't tested them).  So setting a constant D50 for a reservoir is probably not very realistic, but it may be the only real option at the moment.  Your model may have to compromise, overestimating sediment trapping during lower-flow years and underestimating trapping during higher-flow years. 

I might add that my watershed has many, many lakes, and many of these are in sequence, so by the time the particles get to the lower lakes, only the fine particles are left (except what may have eroded from channels).  So my particle sizes can be pretty small on average.  Your reservoir may be more directly connected to sediment sources with a coarser D50. 

Good luck,

-- Jim

---

From: "fatemeh babakhani" <babakhan...@gmail.com>
To: "ArcSWAT" <arc...@googlegroups.com>
Cc: "nbalaji iitm" <nbalaj...@gmail.com>, "SWAT-user" <swat...@googlegroups.com>, "SWAT-CUP" <swat...@googlegroups.com>
Sent: Tuesday, May 31, 2016 11:24:52 AM
Subject: Re: [ArcSWAT:8151] Re: How to estimate sediment from Total Suspended Solids?

Visit this group at https://groups.google.com/group/arcswat.

For more options, visit https://groups.google.com/d/optout.

---

Spam
Not spam
Forget previous vote